My journey began with a deep curiosity for data and technology, and has grown into building AI systems that solve real-world problems and drive meaningful impact.

From developing NL2SQL systems to creating AI-powered research tools, I thrive at the intersection of research and practical implementation. My approach: Research → Build → Deploy.

2,200+ Downloads
Fine-tuned models on HuggingFace
IEEE Publication
Research in personalized AI systems

Expertise

Python & PyTorch LangChain & RAG Large Language Models Natural Language Processing SQL & Database Design CI/CD Pipelines Graph Databases Cloud Deployment

Software Engineer — Gen AI & Data Science

Yes Securities, Mumbai
Aug 2025 - Present

Building scalable data infrastructure with Apache Superset, enabling cross-database visualization and streamlining BI workflows.

Apache Superset Data Visualization Analytics

AI Engineer Intern

Electronic Payment and Services
Jan - Jul 2025

Built NL2SQL agent (90% faster analysis), Jenkins CI/CD pipelines (80% less manual work), and multi-agent automation systems.

NL2SQL LLMs CI/CD

IT Intern

Yes Securities India
May - Jul 2024

Prototyped Neo4j graph databases for financial systems and built RAG-powered document Q&A application.

Neo4j RAG

Volunteer

Being Animal Foundation
May - Jun 2022

Applied technical skills to improve organizational workflows and demonstrate technology's social impact potential.

Social Impact
02
Published Research

Wellness AI

Personalized Health Intelligence

A mobile health assistant that understands individual needs. Fine-tuned Llama 3 model processes real-time wearable data and multiple user metrics to generate personalized diet and fitness recommendations. Published in IEEE, with models serving 2,200+ downloads on HuggingFace.

2,200+
Model Downloads
5+
Tracked Metrics
Llama 3 Model Fine-tuning HuggingFace Mobile AI
03
Live

Retail Sales Predictor

Time Series Forecasting

A forecasting tool that makes time series analysis accessible. Using auto-ARIMA for intelligent model selection, the platform handles datasets up to 200MB and delivers accurate sales predictions through an intuitive interface deployed on Streamlit Cloud.

ARIMA Time Series Analysis Streamlit Statistical Modeling