I am an AI/ML Engineer with strong backend development expertise, focused on building intelligent, scalable, and production-ready systems. My work involves developing machine learning models, designing robust backend services, implementing computer vision pipelines, and deploying AI solutions efficiently.
I am particularly interested in applied AI systems, multimodal architectures, and end-to-end ML pipelines that bridge research and real-world production.
- Deep Learning and Neural Networks
- Transformers and Large Language Models (LLMs)
- Computer Vision (Detection, Tracking, Image Processing)
- Natural Language Processing
- Generative AI and Model Fine-Tuning
- Model Optimization and Inference Efficiency
Libraries & Frameworks:
Python, PyTorch, TensorFlow, OpenCV, Hugging Face, Scikit-learn
- RESTful API Design and Development
- Microservices Architecture
- Authentication and Authorization (JWT, RBAC)
- Database Design and Query Optimization
- Caching and Asynchronous Task Processing
- WebSockets and Real-Time Systems
Technologies:
Django, FastAPI, Flask, Node.js, Redis, MongoDB, PostgreSQL
- Containerization and Deployment
- CI/CD Pipelines
- Cloud Infrastructure Fundamentals
- Linux-Based Development Environments
Tools:
AWS, Docker, Kubernetes, Git, Linux
-
Vision Transformer Implementation
Designed and trained a custom Vision Transformer model for image classification tasks. -
LLM Fine-Tuning Pipeline
Fine-tuned open-source large language models for domain-specific applications, focusing on prompt design and embedding optimization. -
Real-Time Object Detection and Tracking
Implemented and optimized YOLO-based pipelines for real-time inference and edge deployment.
-
Scalable API Service
Built a FastAPI-based microservice architecture with JWT authentication and role-based access control. -
AI Model Serving Platform
Developed a Django backend for deploying, managing, and serving machine learning models in production. -
High-Performance Data Processing Pipeline
Designed ETL workflows using Redis-based task queues for efficient data ingestion and processing.