Senior Python Developer & AI Engineer | LLM-powered Data Platforms
5+ years building scalable cloud-native AI platforms.
Expert in Python · FastAPI · LangGraph · Text2SQL · Agentic Systems.
- Languages: Python • SQL • JavaScript • Java • Bash
- AI / LLM Stack: LangChain • LangGraph • LLMs • Text2SQL Evaluation • Prompt Engineering • AWS AutoGluon
- Backend: FastAPI • REST APIs • Microservices
- Data Engineering: PySpark • Delta Lake • Airflow • Databricks • ETL & Streaming
- Cloud: AWS • GCP
- Databases: MongoDB • Oracle • PostgreSQL • MySQL
Recent Focus: Building and evaluating Text2SQL systems for complex enterprise schemas.
Lead Python Developer
Virtusa (Consulting for Citi Bank) — Jun 2024 – Present
- Built Text2SQL Benchmark Framework — Designed and implemented a comprehensive evaluation system using EX (Execution Accuracy) and QAS (Query Affinity Score) metrics to rigorously test and optimize LLM performance on complex enterprise database schemas.
- Optimized schema context via intelligent pruning and selective retrieval — reduced token consumption and inference costs by 50% while preserving query accuracy.
- Developed LLM-powered agentic workflows using LangGraph + LangChain for dynamic data model generation and conversational interactions.
- Designed FastAPI microservices for metadata extraction, validation, and loading from XLSX + database sources into MongoDB & Oracle.
- Integrated real-time ML training & inference with AWS AutoGluon.
Specialist Programmer
Infosys — May 2019 – Dec 2021
- Engineered real-time streaming pipeline on GCP (Pub/Sub + Cloud Run) resulting in 15% sales increase.
- Built and optimized 4 large-scale ETL pipelines using PySpark, Delta Lake, Airflow & Databricks.
- Achieved 30% reduction in Spark job runtime and 15% lower memory usage.
- M.S. Computer Science — Northeastern University (2022 – 2023)
Teaching & Research Assistant - B.Tech Computer Science — Amrita School of Engineering (2015 – 2019)
- Text2SQL Benchmark Framework (Current) → Open-source style evaluation suite with hybrid metrics (QAS + VAS + LLM) for enterprise Text-to-SQL systems.
- Wander Finds (Gemini AI Hackathon 2024) — Personalized travel recommendations using Gemini + LangChain.
- Shark Tank Analysis & Hate Speech Detection (In-browser ML).
📍 Brandon, FL
📧 kasivisu3109@gmail.com
Building production-grade AI systems that deliver measurable impact.