Sai Krishna Chittanuri

About

ML/NLP engineer with experience at Charles Schwab building an NLP + graph (Neo4j) entity-resolution system to unify fragmented user/device/transaction data and improve downstream model quality and analyst querying. Has applied embeddings (SentenceTransformers + FAISS) with domain fine-tuning to boost hard-case matching recall by ~12% while maintaining precision, and has a track record of hardening scalable Python/Spark pipelines and productionizing fraud models via A/B tests and shadow-mode monitoring.

Experience

Education

Key Strengths

• Built NLP-driven entity resolution pipeline to unify millions of semi-structured behavioral/device/transactional records
• Graph-based data linking with Neo4j to reduce noise and improve downstream ML precision/recall
• Rigorous validation using edge-case sampling, precision/recall checks, and downstream model lift
• Improved hard-case entity matching via SentenceTransformer embeddings + FAISS and domain fine-tuning (recall +12% without precision loss)
• Production-grade data workflow design (modular layers, tests, typing, scheduling with Airflow/AWS Glue, CI/CD, data contracts)
• Scaled a pandas job from 45 minutes to under 5 minutes by refactoring to Spark with partitioning
• Strong model promotion discipline using A/B tests, shadow mode, drift/latency/SLA criteria, and stability gates