Chavvi Bhatia

Migrated legacy codebases into a modular Turborepo monorepo architecture and set up automated deployments to Vercel using GitHub Actions. Built a custom PDF export tool converting Markdown to styled HTML and dynamically rendering PDFs. Developed a tool for dynamic few-shot prompting leveraging Anthropic’s Claude API and Hugging Face Transformers to semantically filter project insights, compute embeddings, and rank content. Added dynamic task-based weightings to prioritize marketing-aligned examples. Created an image generation tool integrated with Replicate’s Flux models and deployed on Vercel. Deployed containerized agent APIs on Google Cloud Platform’s Cloud Run. Built knowledge graphs in Neo4j by transforming relational data from Neon to enable semantic retrieval and reasoning for AI agents.

Designed and implemented an advanced QA analysis tool with ReactJS UI and Python backend hosted on AWS ECS to improve testing efficiency. Led migration of an internal app to AWS cloud for scalability and reliability. Developed a Python data tracing tool for real-time monitoring across S3, Delta Lake, and Snowflake improving data pipeline accuracy. Contributed to chatbot development using Hugging Face for internal mobility optimization. Developed Java-based scripts generating and validating large transaction datasets increasing data integrity checks by 30%. Collaborated with QA teams for data pipeline testing to ensure stability. Built a smart bot leveraging transformers during a hackathon, increasing issue resolution speed by 40%. Led creation of Critical Reconciliation Engine for real-time data validation across client datasets. Enhanced transaction processing utility and automated data reconciliation workflows using Selenium and Cucumber framework. Monitored and resolved regression p

Contributed to reliability and data integrity in financial workflows: improved issue resolution speed by 40% and SRE productivity by 25% during a hackathon via a transformer-based smart ticketing bot; validated 100K+ transactions in IST/UAT with Java to improve data integrity by 30%; spearheaded a Critical Reconciliation Engine reducing manual cross-checks by 40%; boosted QA efficiency with a ReactJS frontend and Python backend on AWS ECS, surfacing 50+ daily issues with downloadable reports; migrated an internal app to AWS for improved scalability and reduced downtime; implemented a Python data-tracing tool across S3, Delta Lake, Snowflake for real-time monitoring, raising pipeline accuracy by 35%; built a Hugging Face chatbot with resume-ranking to streamline HR reviews by 25%; enhanced QA scripting in Java to reduce production errors by 20%; automated data movement with Selenium and Cucumber to reduce manual validation; improved daily regression pipeline uptime to 99%.