cv

Software Engineer, AI Strategy & Solutions

Insight Global, LLC

Leading AI integration and delivery across Universal Creative and Universal Destinations and Experiences, partnering with stakeholders to scope use cases and drive prototypes into productionShipped multiple GenAI solutions, including a domain expert chatbot for internal knowledge discovery and decision support across teamsBuilt end-to-end RAG + agent workflows (ingestion, chunking/embeddings, vector retrieval, grounded generation) with guardrails and source attribution to improve reliability and trustAutomated recurring manual workflows using LLM-based extraction/classification and integrations with enterprise systems, reducing operational overhead and cycle time for high-volume processesImplemented an LLM evaluation & regression harness (golden sets, retrieval checks, prompt/embedding experiments) to prevent quality drift and enable safe iterationHardened solutions for production with secure access controls, logging/monitoring, and documentation to support adoption and long-term maintainability

• GenAI
• RAG
• LLM Agents

Data Scientist

Insight Global, LLC

Spearheaded the integration of AI solutions into media post-production pipelines for FOX Sports, enhancing efficiency and significantly reducing manual effort in post-production workflowsDecreased issue resolution time 35% by developing a Slack AI agent (AWS Bedrock, Lambda) to monitor ML pipelines; added function calling & guardrails to surface drift alerts, exceptions, and health insights in real timeReduced ad-hoc reporting effort 40% by building a RAG conversational assistant (FAISS, Bedrock) that translated natural language to SQL; delivered p95 ~1.3s latency for producersBuilt an LLM evaluation harness (synthetic queries, oracle sets, Ragas precision@k) to tune prompts/embeddings; improved top-5 retrieval by +12 p.p.Optimized serving with prompt caching and lightweight routing on Databricks; reduced per-request cost 28% without quality lossStreamlined operations on 2M+ digital assets by automating metadata extraction and ETL (BigQuery, Spark, Cloud Functions); features powered engagement forecasting (LSTM, transformers) that lifted user retention 25%

• AI/ML
• Data Science
• FOX Sports

Software Engineer

Sakar Robotics

Implemented stereo depth + point-cloud registration (OpenCV, Open3D, ICP) for grasping; increased pick success 18% on cluttered binsPerformed multi-camera calibration (intrinsics/extrinsics, PnP/RANSAC); reduced pose error 22% vs. baselineFused frames into a TSDF map for robust 6-DoF tracking; cut re-localization time during occlusions by 35%Saved $10,000/year by deploying a vision-based construction-site monitoring solution using NeRFs, reducing manual inspection costs and improving safety oversightAchieved 30% precision gains in manipulation by developing reinforcement learning (DDPG + HER), accelerating development cycles

• Computer Vision
• 3D Vision
• Robotics

Mentor

First Lego League (FLL)

Helping middle school students design and code robots

Proprietor

Sai Classes, Pune

Founded and taught the computer science and mathematics curriculum

Recruitment & Retention Manager

Department of Transportation Services, University of Maryland

Master of Engineering

University of Maryland, College Park

Robotics

• Machine Learning
• Deep Learning
• Reinforcement Learning
• Computer Vision