Enterprise RAG
BRIDGE: Multi-Agent Search
Multi-agent RAG platform achieving 93% MRR grounding accuracy using hybrid dense + sparse retrieval for catalog search and knowledge synthesis.
It all started when I took a Machine Learning Course in my undergrad. That opened my eyes to the power of data and algorithms how a few lines of code could reveal hidden patterns, make predictions, and provide actionable insights. What started as curiosity soon grew into a passion for understanding how ML could transform industries and improve decision-making.
After earning my Bachelor of Engineering in Computer Science, I joined Ideapoke Technologies as a Machine Learning Engineer. I worked on large-scale text mining, in-house NER, and retrieval-augmented generation (RAG) pipelines using LangChain and GPT-3.5. A key contribution was optimizing the preprocessing pipeline of the Signalz product, which improved accuracy by 40% and reduced processing time by 60%.
To deepen my research capability, I completed my Master's in Data Science at George Washington University, where I deep-dived into NLP, Reinforcement Learning, and Big Data. As a Graduate Research Assistant on the LAiSER Project (and Recipient of the Global Leaders Scholarship), I built framework APIs on AWS and mapped skills to standards like ESCO and O*NET using Neo4j knowledge graphs. Presenting LAiSER at the 2025 Badge Summit Conference was a true milestone.
Alongside my studies, hackathons became my sandbox. From building real-time coding games at HoyaHacks to co-developing GearGuide (a graph-powered mechanic troubleshooting chatbot using Neo4j and dense+sparse vector search), I loved proving that I can build functional, state-of-the-art AI prototypes on tight timelines. Today, I build production-ready ML pipelines, causal inference analysis, and agentic RAG setups that bridge business value and technical excellence.