Our Portfolio
Explore 28+ production AI projects spanning computer vision, generative AI, robotics, and edge deployment — from hospital operating rooms to autonomous robots.
All Projects
Filter by category to find relevant work
Accident Detector
Real-time accident detection from video using InternVideo2 + BERT.
Badminton Analysis (MatchFlow)
Rally detection, highlights generation, and Streamlit dashboard for badminton match analysis.
Bottle Contamination Detection
Industrial quality control system for detecting contaminated bottles.
Emotion Detection
Real-time emotion detection on Jetson Orin using DeepStream and TAO models.
Facial Landmark Detection
80-point facial landmark detection at 57 FPS on Jetson using DeepStream.
Gesture Detection
Hand gesture classification on Jetson using DeepStream.
Golf Analytics
Swing analysis with pose estimation and club tracking.
People Counting
Occupancy analytics with Kafka cloud integration using DeepStream.
Pose Classification
3D pose estimation on Jetson using DeepStream.
Gun Detector
YOLOv8 gun detection with INT8 quantization deployed on Jetson via DeepStream.
Automatic Number Plate Recognition (ANPR)
License plate recognition system supporting US and Chinese plates.
YOLOv8 Applications Suite
13 modules including car counting, speed estimation, heatmap, segmentation, workout monitor, and more.
Fish Size Calculation
Real-world fish measurement using depth estimation.
AD Blocker (BLOKTV)
TV ad detection using OCR across streaming platforms.
Agentic RAG
5-node workflow with document grading and query rewriting using LangGraph.
Live Video Event Alert
VLM-based real-time video monitoring and alerting system.
LLM with MCP Server
Multi-provider LangGraph + MCP integration supporting Anthropic, OpenAI, and Google.
Open-Vocabulary Object Detection
Zero-shot object detection with TensorRT optimization on Jetson.
Open-Vocabulary Object Segmentation
Zero-shot object segmentation on Jetson using OWL-ViT and MobileSAM.
Semantic Image Search
Multimodal image search with vector database using NV-CLIP and Milvus.
Structured Text Extraction
Document field extraction with VLM pipeline using NVIDIA NIMs.
Travel Chatbot
Conversational travel assistant with human-in-the-loop using LangGraph and Claude.
Course Materials RAG
Full-stack RAG system for course content question and answer.