Souradip Mukherjee

Aspiring AI/ML Engineer with hands-on experience in building LLM-powered applications, RAG pipelines, and fine-tuned models. Proven ability to develop end-to-end AI systems for code automation and enterprise data analysis, improving efficiency and accuracy. Strong foundation in Python, LangChain, FastAPI, and scalable AI deployment.

Internship & Full-time - VITWO

(2025-03 - 2026-04)

Built an AI-Driven Code Testing Framework and a fine-tuned model assistant for sales data

• Developed an AI-powered testing framework using RAG and LangChain to automate bug detection, classification, and fix generation.
• Integrated with GitHub repositories to retrieve relevant code context, enabling precise line-level issue identification.
• Designed context-aware fix suggestions, reducing dependency on manual debugging workflows.
• Reduced debugging time by 70–75% and decreased post-release defects by 20–40%, improving release efficiency.
• Developed a domain-specific fine-tuned Llama-3.1-8B model for sales (O2C) and procurement (P2P) data analysis.
• Built an end-to-end synthetic data pipeline using Groq LLM (Gemma2-9B) and performed QLoRA fine-tuning with Unsloth.
• Enabled natural language querying for financial analysis, trend comparison, and temporal insights.
• Deployed as a FastAPI service with Ollama, enabling real-time business insights.
• Improved answer accuracy by 82% and reduced manual data analysis time by 70%.

B.Tech - Computer Science and Engineering (Data Science) - Future Institute of Engineering and Management, Kolkata, India (2024)

Class XII - Kendriya Vidyalaya Fort William, Kolkata, India