About
Data engineer/scientist and quantitative researcher with 18 years building systems that extract signal from noise.
What I do for clients: I help APRA-regulated risk and data teams build continuous AI assurance tooling — the validation pipelines, drift and calibration monitoring, and audit trails that let ML and LLM systems pass a CPG 235 / BCBS 239 review instead of just a demo. I’m a GARP-certified Risk and AI (RAI) practitioner, so the assurance side is a credential, not a claim. I also deliver end-to-end AI/ML projects and quantitative research, the latter grounded in CFA Level 1 fundamentals. If any of that is your problem, get in touch.
APRA’s April 2026 letter now expects second-line functions to have tooling to independently assess AI systems. That tooling is what I build.
My work spans system engineering, ML, data engineering, portfolio analytics, and economic research—always anchored in first-principles thinking and empirical validation. I don’t chase narratives that don’t survive contact with data.
Current Work
Data Engineering & Analytics
Leading data platform modernization at Macquarie Group: DBT/Iceberg/Redshift architectures, data governance through DBT & Collibra, and observable data pipelines.
Investment Research
Developing quantitative frameworks for portfolio construction and long-term portfolio strategy.
Systems Thinking
Building graph-based intelligence platforms for venture capital analysis and exploring applications of ML/AI to financial domains.
Background
Technical: Data engineering at scale • DBT, Spark, Redshift, Iceberg • ML/AI (graph networks, probabilistic models) • LLMs & Beyond (JEPA) • Modern development (Python, R, AWS) • Blockchain/decentralized systems
Analytical: Quantitative finance • Portfolio analytics • Economic time-series analysis • GARP Risk and AI (RAI) • CFA Level 1 • Data governance (BCBS 239, CPG 235)
Domain Experience: Financial services • Media/publishing • Healthcare • Renewables/energy • Gaming
Previous Roles
- Senior Manager/Lead Engineer at Boson Research Advisory (2023-2025)
- Product Manager at Hydroverse
- Various roles at Sky News, Fairfax Media, Zynga, Oracle, and startups across Sydney, San Francisco, and Bangalore.
Published Work
Crusade Against Kidney Disease (2021)
Research synthesis on slowing CKD progression through evidence-based interventions. Written after watching family members navigate clinical treatment failures. Not-for-profit.
Interests
Outside of work: evolutionary economics, self-sustainability, genetic-based health optimization, and long-term systems thinking.
Work With Me
I take on a small number of engagements per year. The ones that fit best:
- AI assurance & model governance — continuous validation, drift and calibration monitoring, and audit trails built to satisfy CPG 235 / BCBS 239 reviewers.
- End-to-end AI/ML delivery — scoping, data, modelling, and deployment, owned through to a system that runs in production and stays maintainable.
- Quantitative research — portfolio analytics, economic time-series, and financial modelling, anchored in CFA Level 1 fundamentals and delivered as reproducible systems, not throwaway notebooks.
- Data platform modernisation — DBT / Iceberg / Redshift architectures with lineage, governance, and observability as first-class concerns.
If you have models in production and a regulator asking for evidence they behave, or a research question that deserves a real system behind it, that’s exactly the work I do.
Email: bose.debasish@gmail.com
Site Philosophy
This site archives my research, notebooks, and projects. It’s organized for reference and discovery, not chronology. Everything here prioritizes:
- Quantitative reasoning where possible
- First-principles thinking where necessary
- Skepticism toward narratives unsupported by data
The notebooks are working analysis with code and assumptions made explicit — that transparency is the point, and it’s the same rigour I bring to client work.