Mar 2024 – Present · 2 yrs

• • Architecting and implementing an LLM-based AI Agent infrastructure at Everstage, leveraging Langchain for model integration, Typesense vector database for efficient data retrieval, creating a robust and flexible framework to power diverse AI-driven applications within the company's ecosystem.

Oct 2023 – Mar 2024 · 5 mos

• Enhanced a legacy Abstract Syntax Tree (AST) expression evaluation engine by migrating from iterative dataset processing to high-performance Vectorized Pandas operations.
• Developed a Chrome extension that automates deal analysis in Gmail, displaying potential commissions for sales emails from prospective clients.

Jul 2022 – Oct 2023 · 1 yr 3 mos

• Rewrote an existing ETL module, replacing the in-memory pandas implementation with Snowflake, enabling efficient management of massive datasets (10k to billions) and facilitating seamless onboarding of large enterprise customers.
• Achieved a 75% improvement in the ETL pipeline’s runtime by spearheading the implementation of customer data sharding in Snowflake, that resolved delays caused by one client’s queries waiting on locks held by another, reducing the processing time from 12 hours to 3 hours.
• Conceptualized and implemented an incremental ETL pipeline with a change data capture mechanism to optimize data operations (join, union, group by) by avoiding redundant computations and processing only modified records.