data & ai platforms · regulated finance · Gurugram, India

Prashant Singh

Lead AI / Data Engineer — GenAI, ML & Data Platforms

I build and run data and AI platforms for regulated finance. Twelve years across banking and asset management — currently owning curated datasets and regulatory disclosure pipelines at Nomura Asset Management, and prototyping GenAI capabilities that hold up to audit: LLM integration and evaluation, RAG, and agentic workflows with LangGraph and MCP. Deep in Python, Snowflake, Airflow, Spark, and AWS; hands-on ML across credit scoring and fraud detection.

01Skills

Generative AI & LLM

  • Agentic AI (LangGraph, MCP)
  • RAG & Vector Search (ChromaDB, BGE)
  • LLM Integration & Evaluation
  • Prompt Engineering & Guardrails
  • Knowledge Graphs (Neo4j)

Bankopedia's RBI circular summarizer runs as a LangGraph agent over a RAG pipeline. See the project

Machine Learning

  • Predictive Modeling (credit, fraud)
  • scikit-learn / XGBoost
  • Feature Engineering & Evaluation
  • PyTorch

Built credit-scoring and fraud models at Indian Overseas Bank; now writing that up as a series. Financial ML from zero

Languages & Frameworks

  • Python
  • SQL
  • Java
  • PySpark / Scala
  • FastAPI / Django

Six Python packages on PyPI, from NSE market data to a quantitative-finance library. Open source

Data Platforms & Pipelines

  • Snowflake (Snowpipe, Cortex, Dynamic Tables)
  • Apache Spark
  • Databricks
  • dbt
  • Kafka

Built a bank's enterprise Snowflake warehouse; wrote the Spark UI walkthrough my team onboards with. How to read the Spark UI

Orchestration & Cloud

  • Apache Airflow
  • AWS (Glue, EMR, MWAA, Lambda)
  • Docker & CI/CD
  • Temporal

Migrated Control-M and AutoSys batch workloads to Airflow at Macquarie. The migration patterns that held up

Databases & Storage

  • PostgreSQL
  • SQL Server / T-SQL
  • Redis (caching, pub/sub)

Warehouse procs with an explicit error path: TRY…CATCH, ROLLBACK, THROW. The T-SQL post

02Experience

Nomura Asset Management International

Dec 2025 — Present

Lead Software Engineer

Own curated data products and regulatory disclosure pipelines for US funds; prototyping AI and analytics automation with Snowflake Cortex ML and Dynamic Tables.

Macquarie Asset Management

Oct 2024 — Dec 2025

Senior Data Engineer, Assistant Manager

Migrated legacy batch workflows from AutoSys and Control-M to Airflow; evaluated LLM, RAG, and MCP patterns for workflow automation.

Indian Overseas Bank

Jun 2014 — Oct 2024

Python Developer → Senior Data Engineer, Manager

Built credit-scoring and fraud ML models, AML analytics pipelines, and the bank's enterprise Snowflake warehouse; led a six-member engineering team.

03AI Projects

An agentic AI application that ingests Reserve Bank of India circulars and publishes Claude-generated summaries on bankopedia.co.in. Orchestrated as a LangGraph agent workflow over a RAG pipeline — document chunking, BGE embeddings stored in ChromaDB — with prompt engineering, output validation, and guardrails keeping every summary faithful to the source regulatory text.

LangGraphAnthropic ClaudeRAGChromaDBBGE embeddings

Project Synapse

Knowledge Graph

A knowledge graph of a codebase in Neo4j, modeling code structure and relationships, that translates a production error directly into the pinpointed source of the issue and a suggested fix. Combines graph traversal with semantic retrieval over BGE embeddings in ChromaDB, linking error signatures to the relevant code and fix context.

Neo4jChromaDBBGE embeddingsPython

04Open Source

Python packages published on PyPI — market data, quantitative finance, and tooling for regulated-AI workflows.

05Certifications

Course certificates

06Education

07Writing

Notes from production — all 11 posts →

08Get in Touch

Whether it's a role, a data-platform problem, or a question about something I wrote — drop me a line and it lands straight in my inbox.