Developed an AI-Powered Code Improver as a compiler design lab project, merging automation, retrieval, and reasoning to deliver actionable, explainable, and intelligent code enhancement.

• Designed a custom compiler design lab workflow using n8n for automation and advanced LLMs for code optimization.
• Integrated Retrieval-Augmented Generation (RAG) to provide contextual memory, allowing the model to recall project architectures, custom libraries, and coding rules.
• Built a scalable and developer-friendly system that evaluates code performance, safety, and readability, outputting structured enhancement suggestions.

An AI-driven assistant designed to analyze complex government tender documents, extract critical requirements, and eliminate compliance ambiguities.

• Engineered advanced PDF parsing pipelines to convert messy unstructured OCR text from multi-page government tenders into structured JSON formats using Mistral OCR.
• Developed contextual verification layers and custom prompts using LLMs to eliminate hallucination of legal and contractual conditions.
• Created a unified parser architecture that standardizes eligibility, scope, and financial analysis across various department formats.

Contributed to multiple advanced AI chatbot and automation projects at Brahmaastra.ai, streamlining operations and boosting lead generation.

• Developed an AI-powered chatbot for the Homothon Real Estate Expo, handling thousands of user queries about properties and events.
• Built automated WhatsApp Web bots to streamline communication with brokers, clients, and internal team members.
• Designed scalable custom backend workflows and scraping scripts to automate lead extraction and data synchronization across systems.