Projects

Designed and developed a strategic simulation platform in collaboration with the National Defense University to model disease outbreaks and supply chain disruptions across countries.
The tool empowered adjudicators to conduct real-time, turn-based exercises, test response strategies, and generate reports for post-simulation analysis.
This project demonstrated the power of digital simulations in training teams for high-stakes decision-making environments.

Developed a Slack‑integrated AI agent to automate weekly performance tracking for growth and operations teams by detecting KPI anomalies, correlating product rollouts from Jira, and delivering actionable insights.

Designed an interactive Tableau story analyzing gender disparity and crimes against women across Indian states. Integrated demographic, literacy, and population data with crime statistics to uncover regional inequalities and post-2012 reporting shifts.

Engineered a real-time, low-latency system Kafka-based pipeline streaming ratings from UI to a PySpark backend, deploying the app on cloud with a live feedback loop for personalization. Handling 60+ simultaneous ratings with instant recommendations, validating scalability.

Built an AI-powered code similarity detection system to automatically identify duplicate and functionally equivalent code across large repositories.
Using CodeBERT embeddings to represent code semantics and a Siamese neural network for comparison, we trained and evaluated our model on benchmark datasets.
This approach demonstrated how modern NLP techniques can be adapted to software engineering problems, streamlining code review and maintenance.

Engineered a modular AI pipeline for resume classification, job matching, summarization, and interview question generation, leveraging transformer-based language models and semantic similarity to enable scalable, context-aware candidate evaluation.

Engineered an interactive data platform to analyze and visualize federal spending patterns, and designed a custom anomaly scoring framework to identify high-risk fund allocation discrepancies across 30+ agencies; insights were presented to USAspending.gov leadership.

Developed a digital transformation strategy for Starbucks by analyzing its current digital landscape, including mobile app adoption, AI-driven personalization (Deep Brew), and global supply chain systems.
Applied frameworks such as business outcomes mapping, reimagined process flows, and a three-phase transformation roadmap.
Proposed initiatives including Smart Queue management, blockchain-enabled supply chain transparency, AI-driven workforce scheduling, and regional digital expansion, enhancing engagement, operational efficiency, and scalability.

Built a synchronized group-watch platform with WebSockets for real-time video/audio sync across multiple users. Developed custom APIs for session management and recommendations using Microsoft Custom Vision AI. Achieved 3rd place out of 64 teams at the Dish-a-thon competition, demonstrating strong potential for collaborative media streaming.

Developed an AI-powered trading system using RL (Q-learning, OpenAI Gym, TensorTrade) to simulate buy/sell/hold strategies. Explored how RL agents adapt to volatile markets and benchmarked model profitability, demonstrating the potential of AI-driven trading systems.