Pretrained alexnet tensorflow jobs
...layering in additional modules such as automated email campaigns, an analytics dashboard or even a chatbot assistant, provided they plug into the same ML backbone without bloating the codebase. Feel free to suggest alternatives if you have a better architectural approach. Tech stack is flexible as long as you choose tools proven for production-grade AI apps (e.g., Python, FastAPI, Node, React, TensorFlow/PyTorch, PostgreSQL). Code needs to be well-documented and checked into a private Git repository with clear setup instructions so my in-house team can continue iterating after hand-off. Acceptance criteria: 1. The platform trains and deploys ML models on real or sample legal-marketing data and surfaces actionable recommendations. 2. A user with no coding background can crea...
I want to launch a conversational practice bot that feels like a friendly language partner rather than a rigid tutor. The core flow starts with the bot itself proposing everyday scenarios—ordering a cup of chai at a Connaug...Dialogflow CX, or a custom NLP pipeline in Python or Node.js—so long as: • the scenario engine can expand easily with new role-plays, • Hinglish code-switch detection is accurate enough to nudge, not scold, • quizzes follow a true SRS algorithm (SM-2 or similar), and • voice and text remain in sync across web and mobile. Please include a brief outline of your proposed architecture, any pretrained language models or Hindi ASR/TTS services you would leverage, and a timeline for an MVP that covers at least three role-play sc...
...Extract relevant entities (names, dates, IDs, product references, etc.) from the same messages. You will own the entire machine-learning workflow. That means cleaning and exploring the raw email text, crafting useful features, training and tuning your models, and packaging the final solution behind an API that I can call from our existing back-end. Python is a must, and I’m comfortable with either TensorFlow or PyTorch for the deep-learning components—use whichever lets you move fastest. Traditional techniques with Scikit-learn are welcome wherever they make sense. Because the email stream is live, the models have to run efficiently at scale and be easy to retrain when new data arrives. A clean, well-documented repo, reproducible training scripts, and a straightforw...
...climate-responsive façade control framework Expected Deliverables: Machine learning model (Python / MATLAB / suitable platform) Feature importance analysis Predictive control strategy for shading and ventilation Documentation explaining methodology and results Optional: Conceptual AI façade control system diagram Preferred Skills: Machine Learning / Artificial Intelligence Python (Scikit-learn / TensorFlow / PyTorch) or MATLAB Data analysis and predictive modeling Experience with building or façade performance data is a plus Project Type: Short-term project Timeline: Flexible (2–4 weeks preferred)...
...clear decision to the passenger-facing touchscreen. Core vision logic The heart of the project is a YOLO-based model fine-tuned on our curated luggage set. My primary goal is accurate baggage classification; precise dimensioning and weight checks come next, so the model must fuse vision, distance, and scale readings to flag oversize or overweight items. You are free to optimise in PyTorch or TensorFlow, but the final network should run locally on the kiosk and expose an optional Google Cloud endpoint for batch retraining or overflow processing. Touch-first interface The front end runs on a 15” touchscreen and must walk travellers through multilingual (English/Spanish) boarding-pass validation, a payment screen for excess fees, and a final approval printout/QR. I will ...
...centers on end-to-end model creation: cleaning and preprocessing raw text, selecting the right architecture, training, and evaluating performance against clearly defined metrics. I will supply the text dataset along with the target variable I want predicted. You will decide on the most suitable NLP approach—whether that’s classical techniques with scikit-learn or a deep-learning stack such as TensorFlow or PyTorch—document your reasoning, and implement the solution in clean, well-commented Python. Deliverables: • Reproducible code (Jupyter notebook or .py files) with setup instructions • A trained model saved in a shareable format • Evaluation report outlining methodology, key metrics, strengths, and limitations • Brief hand-off sessi...
...production-grade product, not a toy demo. The long-term engagement is several weeks to months, with competitive compensation. Skills We're Looking For Strong experience with deep learning for face/video generation (GANs, diffusion, NeRF, or similar) Hands-on experience with models like SadTalker, Wav2Lip, MuseTalk, LivePortrait, Thin-Plate Spline, FOMM, or similar Proficiency in Python, PyTorch/TensorFlow Experience with real-time streaming (WebSocket, WebRTC, RTMP, GStreamer) Ability to optimize inference for real-time performance (TensorRT, ONNX, model quantization) Bonus: experience with TTS pipelines (Coqui, Bark, XTTS, ElevenLabs integration) How to Enter Build the POC following the specs above Record your live demo (screen + mic, showing real-time sync) Upload you...
...and interpretability both matter. Here’s what I need from you: • A brief data-exploration notebook that highlights key correlations, missing-value handling, and basic visuals. • Feature engineering tailored to the data’s domain (scaling, encoding, derived metrics, etc.). • At least two supervised algorithms (for example, Gradient Boosting and Random Forest in scikit-learn, or an XGBoost/TensorFlow alternative) trained, cross-validated, and benchmarked. • A concise performance comparison using appropriate regression/classification metrics—whichever fits once you see the target variable. • The final, best-performing model saved in a reusable format (pickle/joblib or SavedModel). • A short read-me that explains: setup step...
...and interpretability both matter. Here’s what I need from you: • A brief data-exploration notebook that highlights key correlations, missing-value handling, and basic visuals. • Feature engineering tailored to the data’s domain (scaling, encoding, derived metrics, etc.). • At least two supervised algorithms (for example, Gradient Boosting and Random Forest in scikit-learn, or an XGBoost/TensorFlow alternative) trained, cross-validated, and benchmarked. • A concise performance comparison using appropriate regression/classification metrics—whichever fits once you see the target variable. • The final, best-performing model saved in a reusable format (pickle/joblib or SavedModel). • A short read-me that explains: setup step...
...(vector analysis). 3. State Transition: The algorithm must seamlessly transition from "Bulge Tracking" (closed eyes) to standard "Pupil/Iris Tracking" (open eyes) within the same 30-second video clip. 4. Mobile Optimization: The final module must be lightweight and optimized for real-time or near-real-time performance on iOS and Android devices. Candidate Profile • Expertise in OpenCV, MediaPipe, TensorFlow Lite, or CoreML. • Proven track record in Image Processing or Eye-Tracking projects. • Strong background in Motion Analysis and extracting signals from noisy video data. • Experience in delivering code ready for mobile integration. Application Instructions Please include the following in your proposal: • Examples of previous work in...
...project centres on building a production-ready medical image -classification pipeline that leverages modern deep-learning techniques. I have a labelled dataset and need end-to-end code that ingests the text, handles cleaning and tokenisation, and trains an accurate classifier. Python is the preferred language; The preprocessing must involve Quantum computing techniques using Pennylane. PyTorch, TensorFlow or another mainstream framework is fine as long as the solution is reproducible and easy to extend. Key deliverables: • Well-commented source code (data loading, model, training loop, evaluation) • Clear instructions to run training on a fresh machine (README or notebook) • Metrics report showing accuracy, precision, recall and F1 on a held-out set • ...
...(Finacle, Oracle, Temenos, or similar) │ ├── Developed reconciliation engines for payments │ ├── Implemented fraud detection systems │ └── Deployed systems in regulated environments (finance, healthcare, government) NICE TO HAVE: ├── Experience with Africa's Talking SMS gateway ├── Knowledge of Levenshtein distance or fuzzy matching algorithms ├── Experience with ML model deployment (TensorFlow, PyTorch) ├── Previous work with microservices architecture ├── Experience with Elasticsearch, Kibana (ELK stack) ├── Familiarity with banking regulations (data protection, PCI-DSS) └── Existing relationships with telco or bank technical teams WHAT WE ALREADY HAVE (The Assets You'll Integrate) We are NOT starting from scratch. The following assets are already bui...
suspicious activity detection using deep learning # Features Real-Time Detection: Monitors live camera feeds and detects suspicious activities instantly. Multi-Class Detection: Classifies activities such as hara...fighting, vandalism, and abuse. Alert System: Sends real-time notifications to authorities upon detecting suspicious activities. Dataset-Based Training: Trained on UCF-Crime, AVENUE Video Dataset, and Violent-Flows for high accuracy. Technologies Used YOLOv8: For real-time object and people detection. CNN-LSTM: For recognizing and classifying human activities. Deep Learning Frameworks: TensorFlow, PyTorch. Backend: Flask/Django (replace with your backend framework). Frontend: Minimalistic web interface for real-time monitoring. Alert System: Email (integrated w...
...lighting at the doorway • Automatic in/out logic that prevents double entries when someone lingers near the camera • Admin dashboard with search, manual override and CSV/XLS export • Installation guide plus brief training so I can manage enrolments myself • Source code or licence details, and clear instructions for future camera or user expansion Feel free to use OpenCV, DeepFace, TensorFlow, AWS Rekognition, or another proven toolset—just explain the choice and any recurring costs. I will supply the RTSP stream from the entrance camera and can provision a small Windows or Linux box on-site if you recommend an on-prem edge device. Once everything is running smoothly and the test group’s times are logging correctly for a full week, I&r...
...must be available both as a responsive web application and as native or cross-platform mobile apps, all drawing from a single codebase whenever practical to streamline future maintenance. Core expectations • Modern, engaging UI/UX that feels consistent across web, iOS and Android. • A secure, multi-tenant architecture so each school’s data remains isolated. • An AI layer (think Python, TensorFlow/-Lite, or similar) that plugs into the learning workflow—for example adaptive content or real-time feedback—without locking us into a specific vendor. • Real-time sync between devices, even on spotty school Wi-Fi, and offline caching for mobile. • Admin dashboard with role-based access, permissions and exportable analytics. • C...
I have a collection of ordinary photos that I want to pass through an AI-powered generator and receive noticeably improved results ...point it to an image and get a refined JPEG back in seconds. The generator should handle any everyday picture, whether it’s a night scene, a portrait, a landscape or something completely different, and consistently lift visual quality in the areas that matter most to me: • Brightness aur contrast • Sharpness aur detail • Colors aur tones Preferred output: JPEG. Feel free to lean on pretrained models such as Real-ESRGAN, Stable Diffusion upscalers, or any custom GAN that fits, as long as the final build is easy to run on a mid-range GPU or even CPU if possible. A straightforward README and a few before/after samples will...
...fast delivery (within 24 hours) - Long-term work possible if successful Scope of Work: - Diagnose deep learning pipeline issues - Fix model execution errors - Debug training / inference workflow - Resolve dependency or environment conflicts - Optimize pipeline stability - Ensure end-to-end execution works correctly - Provide brief documentation of fixes Technical Stack: - Python - PyTorch / TensorFlow - HuggingFace / Transformers - CUDA / GPU acceleration - Docker / Linux environment - API integration & Data preprocessing pipeline Requirements: - Strong experience in Deep Learning production workflows - Experience debugging complex AI pipelines - Comfortable working under urgent timelines and ability to start immediately Timeline: Start: Immediately. Expected turnaround: ...
Looking for a highly skilled retail, merchandising, and pricing Data Scientist with deep expertise in AI, Generative AI, and NLP, who has hands-on experience building the following and can walk through in-depth examples and solutions they have implemented across diverse real-world scenarios to drive scalable retail, merchandi...accuracy. Deployment & Integration: Integrate solutions with applications and data systems via APIs and web services, ensuring scalability and reliability. Research & Development of Emerging Technologies: Stay updated on the latest AI/ML advancements and explore opportunities to incorporate innovations into merchandising and pricing transformation initiatives. Frameworks & Tools: TensorFlow, PyTorch, OpenAI, LangChain, and other mod...
...methodology into clean, reproducible code. The core help I’m after is coding itself—covering the full pipeline from data preprocessing through model training to final evaluation and visualisation. I need datasets, well-documented Python scripts or notebooks that I can run end-to-end on my own machine (or a Colab instance). Expect to work with common libraries such as pandas, NumPy, PyTorch or TensorFlow, Hugging Face Transformers, plus Matplotlib or Seaborn for charts—use whichever combination best suits the objectives while keeping dependencies manageable. Deliverables
 • Data preprocessing module that loads the provided datasets, cleans them, applies any necessary tokenisation and splits them into train/validation/test sets. • Training script tha...
Looking for a highly skilled Data Scientist with deep expertise in AI and Generative AI to lead cutting-edge retail merchandising and pricing analytics initiatives, driving scalable forecasting, optimization, and intelligent ...-Implement optimization models to improve operational efficiency and maximize customer value. Manage the full machine learning lifecycle, including model monitoring, retraining, experiment tracking, and performance evaluation. -Develop and maintain CI/CD pipelines for reliable and scalable deployment of data science solutions. -Work with modern machine learning frameworks and tools such as TensorFlow, PyTorch, OpenAI, and LangChain. -Collaborate with cross-functional teams and integrate solutions with existing applications and data systems via APIs and web...
...common pesticide residues on fresh fruit and vegetable samples. My current lab setup streams CSV files over USB; if you have dealt with other device protocols, feel free to propose an efficient data-capture approach. Core requirements • A clean Python pipeline that parses the spectra, performs any necessary preprocessing (baseline correction, smoothing, normalization), and feeds the data into a TensorFlow model. • A well-documented training notebook + scripts so the model can be re-trained when new pesticides or produce types are added. • (Optional but welcome) a complementary computer-vision module. If you have experience with object detection, segmentation, or classic feature extraction, show me how you would fuse image cues with the spectral output to boo...
...can 1) recognise each image, 2) detect all relevant objects inside it, and 3) optionally segment those objects when that adds value to the overall result. The primary goal is accurate detection and classification, but I’d like the system to be flexible enough to switch on pixel-level segmentation whenever it boosts performance. Here is what I’m after: • A clean Python pipeline (PyTorch or TensorFlow preferred, with OpenCV for preprocessing) that ingests my custom images, trains suitable models, and exposes an easy inference script. • Clear evaluation: confusion matrices, mAP/IoU scores, and plots of loss/accuracy over epochs so I can judge progress at a glance. • A concise, publication-ready report written in LaTeX. I will supply the template; yo...
I need an AI-driven workflow that ingests raw video (MP4/MOV) and automatically follows a chosen object through every frame, delivering a clean alpha-masked output I can drop straight into my edit. The goal is to save me from manual rotoscoping; accuracy and speed are both critical. Here’s how I picture the collaboration: you build or adapt a model—using tools such as PyTorch, TensorFlow, Detectron2, or a proven OpenCV pipeline—that detects the target, keeps the mask tight even during fast motion, and exports either a keyed video or a PNG sequence. A small sample clip will be provided for you to demonstrate proof of concept; once results look solid we’ll run the system on the full batch. Deliverables • Python or command-line script (with all depe...
...AI-powered image-recognition tool that focuses exclusively on identifying everyday household items. The core requirement is clear: when the model sees a photo, it should reliably detect and classify objects such as chairs, tables, kettles, lamps, and similar items that people typically keep at home. Here is how I picture the workflow and final hand-off: • Model: A well-trained neural network (TensorFlow, PyTorch, or a comparable framework) tuned for object detection/classification. • Dataset handling: Either you assemble and label a suitable open-source dataset or guide me on licensing a ready-made one; in either case, the final dataset or clear reproducibility steps must be included. • Inference pipeline: A simple script or lightweight API endpoint so I can ...
...feeds or historical fault libraries later, so a clean, extensible data pipeline matters. Key deliverables • Cross-platform mobile app (iOS + Android) built with a modern stack—React Native, Flutter, or another framework you are comfortable with. • Fault-tree engine that mirrors Airbus AMM logic and lets me update procedures without redeploying the whole app. • Predictive module (Python/TensorFlow, PyTorch, or similar) that ranks probable troubleshooting branches based on past fixes. • Secure local/remote storage of maintenance logs, plus export in CSV or JSON for MIS upload. • Clear documentation and a short video demo showing the workflow on an A320 use-case. Acceptance criteria 1. Given a sample logbook entry “F/CTL PRIM1 FAULT...
...returns its class. • Documentation: concise README explaining setup, dependencies, and how to retrain with fresh data. Acceptance criteria 1. Minimum F1-score of 0.85 on the hold-out test set I will supply. 2. Reproducible environment ( or ). 3. Code delivered via private Git repository or ZIP file. Tools you might consider include Python 3.11, PyTorch or TensorFlow, HuggingFace Transformers, spaCy, and scikit-learn; feel free to propose alternatives if they achieve equal or better results. I will review interim results as soon as you have an initial baseline, then we can iterate on hyper-parameters, class imbalance handling, and deployment details....
...shots, slight rolls, and images with minimal vertical cues. 2. Batch processing and an interactive before-after comparison are mandatory. 3. Front-end must run in any modern browser; a minimal back-end is fine as long as uploads are secure and temporary. 4. Code must be clean and well-documented so I can extend it later. I have no fixed stack preference, so feel free to propose OpenCV, TensorFlow, or any combination of JavaScript (e.g. WebAssembly-compiled OpenCV), Python (FastAPI, Flask), or other technologies you’re comfortable with. Please outline: • The libraries and frameworks you’ll rely on. • Your approach to vanishing-point detection and homography estimation (e.g., RANSAC line clustering, deep-learning refinement, etc.). • Expected ...
...me: • Interactive graphics and videos must steal the show—they should illustrate key concepts, animate data, and invite the viewer to click, pause, or explore. • Avatars must look realistic, lip-sync flawlessly, and be easy to re-skin for future episodes. • The entire pipeline—from text input to final MP4—should run with minimal manual tweaking, whether you build it in Python with PyTorch/TensorFlow or orchestrate commercial generative-media APIs. Deliverables 1. A working proof-of-concept that accepts a text script, generates the realistic avatar narration, and stitches in AI-created graphics/video to produce a cohesive training module. 2. Source code plus clear setup and usage instructions. 3. A short sample episode built from one of my...
...machine-learning model (anomaly detection or sequence-based classification) optimised for on-device or near-edge execution. 2. Implement a decision layer that selects one of three responses—encrypt, quarantine, or alert—based on the model’s confidence score. 3. Provide clean, well-commented Python code (TensorFlow, PyTorch or scikit-learn are all acceptable) plus a short README explaining data preprocessing, hyper-parameters and how to port the model to an embedded runtime (e.g., TensorFlow Lite, ONNX). 4. Supply a small synthetic data set and demonstrate at least 90 % accuracy in distinguishing normal from suspicious activity during a live demo or recorded notebook. Acceptance criteria • Model trains and runs locally on a laptop within 10 min...
...into Reinforcement Learning or NLP later, that flexibility will be a plus for the longer roadmap, but the immediate priority is Deep Learning mastery. The sessions will be delivered live (online or hybrid can be arranged), and I’ll rely on you to: • Shape a clear, week-by-week syllabus covering CNNs, RNNs, transformers, optimisation tricks, model interpretability and MLOps basics using Python, TensorFlow or PyTorch • Provide concise slide decks, hands-on notebooks (Jupyter/Colab) and at least three graded mini-projects that mirror industry use-cases • Guide learners through code reviews and Q&A, then wrap up with a capstone evaluation and feedback report All teaching material must be original or properly licensed, and ready for hand-off at the end ...
...training, validation, and test sets, with any necessary feature engineering you judge appropriate. I have no fixed preference on the final algorithm—linear models, tree ensembles, or a small neural network are all acceptable as long as they deliver solid predictive accuracy and are easy to retrain when I add more data. Please build the solution in standard Python tooling (pandas, scikit-learn, TensorFlow or PyTorch only if the accuracy gains justify it) and present the work in a Jupyter Notebook. Your notebook should walk me through: • data import, preprocessing, and exploratory visuals • model selection and cross-validated performance metrics • prediction of W/L ratio on unseen inputs • a short optimisation routine that searches the design space...
...dosage, batch number, expiry date, manufacturer, and any other legible data. • Return the result programmatically as a JSON object so it can be stored or sent to our API. Accuracy is more important than speed, but I still expect real-time feedback on focus and framing. You’re free to leverage mobile-friendly vision libraries (Google ML Kit, Tesseract, Vision Framework, etc.) or a custom TensorFlow-Lite model if that yields better results. Everything must run on-device; no cloud calls. Deliverables: 1. Full source code for the iOS and Android implementation (native, Flutter, or React Native—use what lets you hit the quality bar fastest). 2. A short read-me that explains build steps, dependencies, and the JSON schema. 3. Sample JSON output from at least t...
...machine learning model that reliably classifies each photo into the correct category. Your job is to design, train, and evaluate the full image-classification pipeline. You may build from scratch or fine-tune a proven architecture such as ResNet, EfficientNet, MobileNet, or a vision transformer—as long as the final model meets the accuracy targets we set together. Feel free to work in PyTorch or TensorFlow/Keras; I’m comfortable deploying either. What I’ll provide • A structured folder of training, validation, and test images • Category labels and a brief data dictionary • Access to a GPU instance if you need it What I need back 1. Clean, well-commented code (Jupyter notebook or Python scripts) that handles preprocessing, augmentation, trai...
...machine-learning model (anomaly detection or sequence-based classification) optimised for on-device or near-edge execution. 2. Implement a decision layer that selects one of three responses—encrypt, quarantine, or alert—based on the model’s confidence score. 3. Provide clean, well-commented Python code (TensorFlow, PyTorch or scikit-learn are all acceptable) plus a short README explaining data preprocessing, hyper-parameters and how to port the model to an embedded runtime (e.g., TensorFlow Lite, ONNX). 4. Supply a small synthetic data set and demonstrate at least 90 % accuracy in distinguishing normal from suspicious activity during a live demo or recorded notebook. Acceptance criteria • Model trains and runs locally on a laptop within 10 min...
I’m building a camera-based system that runs on an NVIDIA Jetson and, in real time, detects faces and recognises emotions. The entire solution must be coded in Python. For face localisation I’d like a fast deep-learning detector—SSD or YOLO—so the frame rate stays smooth on Jetson hardware. Once a face is found, a TensorFlow model should assign an emotion label (happy, sad, angry, surprised, neutral, etc.) together with its confidence score. The video stream has to overlay these results live, log every reading with a timestamp, and trigger a visual or audible alert whenever negative emotions are detected repeatedly within a short window. A lightweight dashboard served with either Streamlit or Flask will let me: • watch the annotated video feed •...
...application Assessment & Gap Analysis: (Optional/Phase 2) AI-driven tests or suggestions for courses to help candidates bridge skill RequirementsFrontend: Modern frameworks such as React, , or Angular for a responsive and intuitive : Scalable environments using Python (Django/Flask), Node.js, or Frameworks: Experience with OpenAI APIs, LangChain, TensorFlow, or PyTorch for building matching models and resume : Secure and efficient data handling with PostgreSQL, MySQL, or : Secure login/registration with role-based access control and optional OTP functional web application (responsive for mobile/desktop).Source code with clean, well-documented with third-party
...configure camera (USB/IP) • Implement real-time face detection • Integrate pre-trained emotion recognition model (no training required) • Display emotion + confidence score live • Log events with timestamps • Build a simple dashboard (Streamlit or Flask) • Add alert logic (e.g. repeated negative emotion → warning) ⸻ Required Skills • NVIDIA Jetson (VERY IMPORTANT) • Python • OpenCV • TensorFlow or PyTorch • Real-time video processing • Linux / Ubuntu • Flask or Streamlit Bonus: • GStreamer / DeepStream experience • Previous edge AI or surveillance projects ⸻ Deliverables • Fully working prototype on Jetson • Clean source code • Installation/setup guide • Dem...
...professional monitoring console: two circular gauges for the intensity and effort scores, plus a central sphere whose colour and animation state change with cumulative findings. A scrollable timeline should let me jump straight to any highlighted event and hear a 10- to 30-second trimmed clip without re-encoding the whole file. Technical freedom is yours—if Python libraries such as librosa, PyTorch or TensorFlow serve the detection, great; if you prefer another stack, convince me. The front end can be React, Vue, or a comparable modern framework; D3, or Plotly can drive the visualisation layer. Deliverables • Trained detection model and reproducible inference script • REST or local API that produces per-second labels and WOB metrics • Web dashboard r...
AI/ML Engineer (7+ Years Experience) J...projects and collaborate with cross-functional teams to deliver scalable AI solutions. Key Responsibilities: Develop and deploy machine learning models Perform data analysis, preprocessing, and feature engineering Work with AI/ML frameworks like TensorFlow, PyTorch, or Scikit-learn Optimize models for performance and scalability Collaborate with teams to deliver project requirements Requirements: 7+ years of experience in AI/ML or Data Science Strong knowledge of Python Experience with Machine Learning, Deep Learning Hands-on experience with TensorFlow / PyTorch Knowledge of NLP, Computer Vision, or similar domains Good problem-solving and communication skills Preferred Locations: Egypt, Philippines, Pakistan, Bangladesh, Indonesia,...
...configure camera (USB/IP) • Implement real-time face detection • Integrate pre-trained emotion recognition model (no training required) • Display emotion + confidence score live • Log events with timestamps • Build a simple dashboard (Streamlit or Flask) • Add alert logic (e.g. repeated negative emotion → warning) ⸻ Required Skills • NVIDIA Jetson (VERY IMPORTANT) • Python • OpenCV • TensorFlow or PyTorch • Real-time video processing • Linux / Ubuntu • Flask or Streamlit Bonus: • GStreamer / DeepStream experience • Previous edge AI or surveillance projects ⸻ Deliverables • Fully working prototype on Jetson • Clean source code • Installation/setup guide • Dem...
...create a user-friendly interface for easy interaction Key Features: Automated task scheduling and execution Natural Language Processing (NLP) for understanding user commands Data analysis and decision-making capabilities Integration with third-party tools (email, databases, apps) Real-time monitoring and reporting Technologies Used: Programming Languages: Python / JavaScript AI Frameworks: TensorFlow / PyTorch APIs: OpenAI API, Google APIs Database: MySQL / MongoDB Tools: Automation platforms like Zapier or custom scripts Working Principle: The system collects input from users (text or voice), processes it using AI models, and then performs the required task automatically. For example, it can read emails, categorize them, send responses, or update records in a database witho...
I need an end-to-end solution that can automatically detect marathon bib numbers from footage captured by two fixed ...upload/edit the bib-to-phone spreadsheet – swap the overlay logo. Deliverables 1. Fully working prototype deployed on a cloud VM or local server, ready for race-day use. 2. Source code with clear documentation and installation script. 3. Admin guide (PDF or short video) showing setup, operation, and troubleshooting. 4. One remote handover session for live Q&A. I’m open to your preferred stack—OpenCV, TensorFlow, PyTorch, FFmpeg, or a commercial vision API—so long as licensing fits an event production environment and the admin UI remains dead-simple. Please outline your proposed architecture, expected accuracy, and any hardware specs I s...
...dependencies. 3. Final raster layers align correctly with standard geographic projections and pass spot-checks against in-situ well data supplied later in the project. When you reply, focus on your experience with satellite hydrology, geospatial machine learning, and any prior work that combined stress-recharge assessments. Briefly outline the toolchain you prefer (e.g., Google Earth Engine, xarray, TensorFlow, scikit-learn) and how quickly you can deliver the first milestone of cleaned input data and an initial baseline model....
...detection) • Audio (spectral patterns, voice anomalies, speaker embeddings) Core Requirements: • Multi-modal deepfake detection (Image, Video, Audio) • Clean and scalable Python-based architecture • Support for datasets like FaceForensics++ and DFDC • Training, evaluation, and inference pipeline • Option to run via CLI, Jupyter Notebook, or executable (.exe) Tech Stack (Preferred): Python, OpenCV, TensorFlow/PyTorch, scikit-learn Audio processing tools (LibROSA, etc.) Frontend + Backend integration (optional but preferred) Additional Features (Bonus): • Web interface (FastAPI + frontend) • Runtime learning / feedback system • Lightweight dataset crawler • Performance optimization for low-resource systems Deliverables: &bul...
...research and implement reinforcement learning algorithms, experiment quickly, tune hyperparameters, and evaluate against clear success metrics. • Integration with existing systems – wrap trained models behind REST/GraphQL endpoints, containerise (Docker/Kubernetes), and wire everything into my current Python micro-services stack on AWS. Everything is Python-first, so fluency with PyTorch or TensorFlow, pandas, NumPy, and popular RL libraries (Stable-Baselines3, Ray RLlib, or similar) is expected. Familiarity with CI/CD (GitHub Actions), infrastructure-as-code, and basic DevOps will make collaboration smoother. Deliverables I’m expecting: 1. Reproducible training pipeline with documented code. 2. Baseline RL model that reaches the agreed-upon performance ben...
...appear only partially. The emphasis is on distribution-robust performance under severe occlusion. I am mainly looking for large models to be used e.g., Mamba / hybrid Transformer models for human pose estimation. The approach I have in mind mixes large Transformer backbones with geometric priors think part-affinity refinements, kinematic graph constraints, or similar. Frameworks such as PyTorch, TensorFlow, Detectron2, or MMPose are all fine as long as the pipeline stays fully reproducible on a single GPU. Deliverables • An occlusion-aware pose estimation model with source code and training scripts • Pre-trained weights plus a clear read-me on how to reproduce results end-to-end • A concise tech report detailing architecture choices, training schedule, metr...
...camera streams, runs fast and accurate inference, and pushes results reliably from edge devices. The current target hardware is NVIDIA Jetson, so every design choice—from model architecture to post-processing—must respect its compute limits while still keeping total end-to-end latency under 200 ms. The core work revolves around training, tuning, and deploying YOLO-style detectors in PyTorch (TensorFlow knowledge is welcome if it helps optimisation). You will refine the models for two challenging scenarios that matter most to our roadside installations: low-light environments and high-speed vehicle movement. Image enhancement, motion-blur compensation, and clever data-augmentation strategies are all fair game as long as they translate into measurable accuracy gains af...
