Open to opportunities

Abhinav Bajpai

@abhinavbajpai2

Message

Software Engineer at CUS Solutions who improved bookkeeping exception accuracy by 17% across 6M+ monthly transactions.

India

Message

At CUS Solutions, I lead Python agent workflows that automate bookkeeping exception handling across 6M+ monthly transactions, improving resolution accuracy by 17%. I also developed hybrid semantic search that reduced policy lookup latency by 29%.

Earlier, I designed Java and Spring Boot services at Tata Consultancy Services for an enterprise platform serving 400,000+ users. My work has also included Azure telemetry pipelines at Meltek Inc. and research agent prototypes at Syracuse University.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer

Current

CUS Solutions

Jun 2026 - Present (5 months)

Led LangGraph-based Python agent workflows for automated bookkeeping exception handling, improving resolution accuracy by 17%. Architected hybrid semantic search with LangChain and OpenSearch, reducing policy lookup latency by 29%.

LangGraph Lang Chain RAG OpenSearch Elasticsearch fastAPI Redis Hugging Face Transformers PyTorch MLFlow Prometheus Evidently AI Python ETL SQL

Research Intern

Syracuse University

Jun 2025 - Jun 2026 (1 year)

Designed and evaluated LangGraph agent prototypes for automated analysis of research papers and source code. Benchmarked retrieval strategies and packaged models behind containerized FastAPI endpoints.

LangGraph fastAPI Docker BM25 Cross Encoder Re Ranking Embeddings Research Python

Software Developer Intern

Meltek Inc.

May 2024 - Aug 2024 (3 months)

Built a real-time telemetry ingestion pipeline on Azure Event Hubs handling 500K+ daily events. Developed FastAPI services with Pydantic schema enforcement, cutting data quality errors by 38%.

Azure Event Hubs fastAPI Pydantic OAuth OIDC Microsoft Entra ID SOC2 Python

Systems Engineer

Tata Consultancy Services

May 2019 - Jul 2023 (4 years 2 months)

Designed and maintained Java/Spring Boot backend services and RESTful APIs for an enterprise platform serving 400,000+ users. Architected PostgreSQL infrastructure with replication and partitioning, improving query performance by 75%.