Welcome! I'm an Electrical Engineering graduate turned Python AI/ML Developer with 1.5 years of hands-on experience at Autobits Labs. I specialize in building intelligent systems, backend microservices, and automation tools from the ground up, with a strong focus on performance and reliability. I thrive on dissecting complex problems and delivering robust, well-researched solutions.

• 🛠️ Building high-performance AI/ML systems (Computer Vision focus).
• 🚀 Designing and implementing backend microservices (FastAPI, Flask, Django).
• ⚡ Optimizing for real-time processing and GPU acceleration (CUDA, ONNX Runtime).
• ☁️ Experienced with containerization (Docker, Docker Compose) and Linux environments (Ubuntu).
• 🌱 Always learning and exploring new technologies in AI, backend development, and cloud platforms.
• 🤝 Open to collaborating on interesting projects and discussing innovative ideas.
• 💼 Seeking challenging opportunities as a Python AI/ML Developer.

• Description: Designed and built a high-performance, scalable platform for real-time face recognition across multiple IP cameras. Features include live monitoring, subject management, access control, and analytics via a web dashboard. (Developed professionally at Autobits Labs - Source code is proprietary).
• Key Features & Concepts:
  • Decoupled microservices architecture (FastAPI Control & Processing services).
  • GPU-accelerated AI (InsightFace/ONNX Runtime) for detection & recognition.
  • Self-healing video streams (FFmpeg/CUDA) with automatic restarts for high availability.
  • Real-time dashboard updates via WebSockets.
  • Asynchronous operations (asyncio, SQLAlchemy, httpx).
  • Comprehensive permission system & background notification service.
  • Fully containerized with Docker and Docker Compose.
• Skills Demonstrated: Backend Development, Microservices, AI/ML Integration, GPU Acceleration, Real-Time Systems, Database Design (PostgreSQL), Docker, Performance Optimization, System Resilience, FFmpeg Pipelines, Asynchronous Programming.

• Description: An application that uses a webcam to detect hand gestures (ISL for 0-9, A-Z) in real-time via MediaPipe, classifies them using a TensorFlow/Keras model, and provides spoken feedback using asynchronous text-to-speech. Features prediction debouncing for stability.
• Technologies: Python, TensorFlow/Keras, MediaPipe, OpenCV, NumPy, threading, queue, pyttsx3.
• Skills Demonstrated: Computer Vision, ML Model Integration, Real-Time Processing, Concurrency (Threading), UI Feedback (Visual & Audio).

• Description: Developed machine learning models (PLSR, Transformers) using TensorFlow to predict pigmentation and capsaicinoid concentrations in chili powder from spectrography data.
• Results: Increased pigmentation prediction accuracy by 20% and capsaicinoid prediction accuracy by 15% compared to baseline models.
• Skills Demonstrated: Machine Learning, Data Analysis (Pandas, NumPy), Model Development (TensorFlow), Regression, Feature Engineering, Spectrography Data Handling.