About Me
Generative AI Engineer and Computer Science graduate (2026) with experience in designing AI agents, LLM applications, and retrieval-augmented generation (RAG) systems for real-world automation and knowledge-intensive tasks. Skilled in building scalable, production-ready AI solutions involving orchestration, fine-tuning, evaluation, and deployment of large language model applications.
Skills & Technologies
Work Experience
Python Developer Intern
Developed Python applications using 10+ core programming concepts and exception handling to improve reliability. Conducted exploratory data analysis and preprocessing on 2,000+ records with Pandas and NumPy, reducing data inconsistencies by 25% and enabling efficient data analysis.
AWS Cloud Intern
Gained practical experience on 5+ AWS services including S3, EC2, IAM, CloudWatch, SPICE and QuickSight. Applied cloud-native workflows to deploy scalable AI and analytics solutions, improving business intelligence reporting and facilitating interactive visualization for 1,000+ data records.
Projects
Agentic Corrective RAG System
Engineered an Agentic Corrective Retrieval Augmented Generation (CRAG) system using Python, LangGraph for orchestration, ChromaDB vector storage, Groq-powered inference, and Tavily web search to dynamically select between local document retrieval and external web sources. Built a Dockerized, Streamlit-based multi-document AI assistant with semantic search and confidence-based retrieval, reducing token usage by 45% and deployed on HuggingFace spaces.
VCoder – Fine-Tuned Python Code Generation Model
Fine-tuned Qwen2.5-Coder-3B instruct model on 15,000+ rows of Python dataset using Transformers, QLoRA, and Unsloth for parameter-efficient adaptation. Benchmarked on 100-task HumanEval subset showing 7-point Pass@1 improvement from 61% to 68%. Deployed a quantized GGUF model using Ollama and published on Hugging Face, attracting 100+ downloads from the open-source community.
Smart Automated Lead Generation & Cold Email AI Agent
Developed an autonomous AI agent workflow with n8n, Python, JSON, Groq, and GCP capable of processing 100+ leads per execution by integrating web scraping, lead enrichment, and AI-powered emails. Automated website analysis and reduced manual lead qualification and outreach time by over 90%, boosting campaign efficiency.
Education
Bachelor of Engineering
Computer Science and Engineering