AUTHOR
Prashant Singh
Lead AI / Data Engineer · GenAI & Machine Learning · 12+ Years in Banking & Asset Management
Prashant is a Lead Software Engineer with 12+ years of experience building and operating production data and AI-enabled platforms in banking and asset management. He spent a decade at Indian Overseas Bank building credit-scoring, fraud-intelligence, and AML analytics systems, and now leads engineering for fund data platforms at Nomura Asset Management International. He built Bankopedia — including its AI-powered RBI circular summarizer — to make practitioner-grade banking knowledge accessible to every banking professional in India.
Career
From a Public Sector Bank's data desk to leading AI-enabled platforms in global asset management.
Education
B.Tech CS · IIM Kashipur EDP
Computer Science graduate (Anand Engineering College, Agra), followed by M.Tech research at MITS Gwalior — published in the International Journal of Computer Applications(2013). Completed IIM Kashipur's Executive Development Program in Applied Financial Risk Management (2025), covering market, credit, and operational risk modeling in Python and R.
Public Sector Banking
Indian Overseas Bank · 10 Years
Rose from Assistant Manager to Manager(2014–2024) building the bank's data and risk systems — credit-scoring and loan-default models, fraud-intelligence and AML monitoring pipelines, Basel II/III reporting controls, and an event-driven loan-origination platform — while leading a six-member engineering team.
Now (2024 – Present)
Global Asset Management
Moved to Macquarie Asset Managementin 2024, and continued through Nomura's acquisition of its public investments business — now Lead Software Engineer at Nomura Asset Management International, owning curated data products for U.S. mutual funds, ETFs, and managed accounts, and prototyping GenAI capabilities under governance controls.
Professional Experience
Twelve years of progressive engineering roles across banking and asset management.
Lead Software Engineer — Nomura Asset Management International
Dec 2025 – Present · Gurugram
Leads end-to-end data and AI-enabled platforms for fund performance, pricing, positions, and regulatory disclosures. Prototypes AI and analytics automation (Snowflake Dynamic Tables, Cortex ML) under validation, security, and governance controls; drives architecture reviews, CI/CD, and mentoring.
Senior Data Engineer — Macquarie Asset Management
Oct 2024 – Dec 2025 · Gurugram
Evaluated LLM, RAG, and MCP-based tool-integration patterns for workflow automation. Migrated legacy EAGLE PACE and STAR batch workflows to Apache Airflow, and built reusable PySpark transformation components with reconciliation and completeness checks.
Senior Data Engineer, Manager — Indian Overseas Bank
Jun 2022 – Oct 2024 · Erode
Built fraud-intelligence, AML monitoring, and near-real-time transaction-alerting pipelines (40% faster detection). Architected an event-driven loan-origination platform on FastAPI, Kafka, Redis, and Temporal. Led a six-member engineering team.
Data Engineer, Manager — Indian Overseas Bank
Jun 2019 – Jun 2022 · Erode
Developed credit-scoring and loan-default prediction models for risk assessment. Designed credit-risk pipelines with Airflow, Spark, and AWS, implemented Basel II/III reporting controls, and architected a governed Snowflake warehouse for retail-banking analytics.
Python Developer, Assistant Manager — Indian Overseas Bank
Jun 2015 – Jun 2019 · Panjim
Built the Outlier Case Alert System for anomaly-based transaction monitoring and automated daily Python and SQL data flows — reducing manual processing while improving auditability and control coverage.
Junior Python Developer, Assistant Manager — Indian Overseas Bank
Jun 2014 – Jun 2015 · Chennai
Developed a risk-based internal-audit management system in Django and MySQL with risk-prioritized scheduling, review, and approval workflows.
Areas of Expertise
Production systems shipped in regulated financial environments — not toy projects.
Generative AI & LLM Engineering
LLM integration and evaluation (Anthropic Claude), RAG over vector databases (ChromaDB, BGE embeddings), agentic workflows with LangGraph, Model Context Protocol (MCP) tool integration, prompt engineering, and output validation with security guardrails.
Machine Learning for Banking Risk
Credit-scoring, loan-default prediction, fraud-scoring, and anomaly-detection models built and deployed in regulated environments — PyTorch, scikit-learn, XGBoost, feature engineering, and model evaluation and monitoring.
Data Engineering & Pipelines
End-to-end batch and streaming pipelines with Apache Spark, Airflow, Kafka, and dbt; Snowflake (Snowpipe, Streams & Tasks, Dynamic Tables, Snowpark); AWS (S3, Glue, Lambda, Redshift, EMR, MWAA); Docker and CI/CD.
Regulatory & Risk Analytics
Basel II/III reporting controls, AML and financial-crime analytics, KYC validation and data governance embedded into production pipelines — backed by IIM Kashipur's Executive Development Program in Applied Financial Risk Management.
Certifications & Credentials
Credentials spanning cloud AI, LLM application engineering, and data-pipeline orchestration.
AWS Certified AI Practitioner
Amazon Web Services
AWS Certified Cloud Practitioner
Amazon Web Services
Build AI Apps with MCP Server
DeepLearning.AI (2026)
LLM Applications with LangChain (RAG)
DataCamp (2026)
JavaScript RAG Web Apps with LlamaIndex
DeepLearning.AI (2026)
Generative AI with LLMs
DeepLearning.AI
DAG Authoring for Apache Airflow 3
Astronomer Certification (2026)
Publication: Neha Agarwal and Prashant, “Incremental Associative Memory Model Algorithm for Highly Scalable Recommender Systems,” International Journal of Computer Applications, vol. 68, no. 21, Apr 2013 · DOI: 10.5120/11705-7317
ON BANKOPEDIA
Why he built this platform
After a decade inside a Public Sector Bank, Prashant saw the same problem repeatedly: banking knowledge was fragmented, RBI circulars were impenetrable for most officers, and exam preparation material was too generic to be useful in practice. Bankopedia is his answer — and its RBI circular summaries are generated by an agentic AI pipeline he engineered: a LangGraph workflow with retrieval-augmented generation over the circular corpus, powered by Anthropic Claude, with guardrails that keep every summary faithful to the source regulatory text.