About Me
AI Engineer with 3+ years of R&D experience specializing in Computer Vision, Edge AI, and Large Language Models. I have a proven track record of architecting end-to-end machine learning pipelines, including a Dockerized …
AI Engineer with 3+ years of R&D experience specializing in Computer Vision, Edge AI, and Large Language Models. I have a proven track record of architecting end-to-end machine learning pipelines, including a Dockerized FinTech fraud detection system that achieved 98% accuracy in real-time document verification. My expertise spans optimizing deep learning models for edge devices like Jetson Orin Nano and developing multimodal LLM agents integrated with sensor data. I am highly skilled in bridging the gap between cutting-edge research and practical deployment using PyTorch, YOLO, TensorRT, and model quantization. As a published researcher, I am passionate about leveraging scalable ML infrastructure to build enterprise-grade applications and drive innovation
Experience
Machine Learning Engineer
Architected an Immigration Document Verification System using Siamese Networks and YOLO, detecting fake passports via UV light pattern analysis with high precision., Engineered a Dockerized deployment pipeline, integrating OCR-free text extraction (Donut) and face comparison models for real-time identity verification., Optimized AI inference using model quantization, enabling efficient on-device performance for mobile and edge use cases., Authored technical research presented at ICGHIT 2025 regarding non-visible-light document verification.
Machine Learning Engineer
Architected an Immigration Document Verification System using Siamese Networks and YOLO, detecting fake passports via UV light pattern analysis with high precision.
Engineered a Dockerized deployment pipeline, integrating OCR-free text extraction (Donut) and face comparison models for real-time identity verification.
Optimized AI inference using model quantization, enabling efficient on-device performance for mobile and edge use cases.
Authored technical research presented at ICGHIT 2025 regarding non-visible-light document verification.
AI Research Engineer
Led the development of “Rumi”, a multimodal LLM agent integrated with EEG sensors for real-time human emotion detection and adaptive response generation.
Built and deployed a full-stack web application using Flask to serve the LLM model, ensuring scalability and real-time user interaction.
Developed a Text-to-Video generation pipeline using GANs, optimizing for high-fidelity output from natural language prompts.
AI Research Engineer
Led the development of ”Rumi”, a multimodal LLM agent integrated with EEG sensors for real-time human emotion detection and adaptive response generation., Built and deployed a full-stack web application using Flask to serve the LLM model, ensuring scalability and real-time user interaction., Developed a Text-to-Video generation pipeline using GANs, optimizing for high-fidelity output from natural language prompts.
PROJECTS
Fake Passport Detection System
Developed an AI-powered passport verification system integrating an optimized YOLO model for document detection and Donut for OCR-free text extraction. Deployed a Siamese UV Sample Matching backend in Docker for authenticity verification under UV light, utilizing model compression for efficient on-device performance.Achieved a 98% accuracy rate in fraud detection
Rumi - Multimodal LLM Agent
Led the development of "Rumi", a multimodal LLM agent integrated with EEG sensors for real-time human emotion detection and adaptive response generation. Built and deployed a full-stack web application using Flask to serve the LLM model, ensuring scalability and real-time user interaction. Developed a Text-to-Video generation pipeline using GANs, optimizing for high-fidelity output from natural language prompts.