Data & AI Advisor for Growing Businesses

Sayan Chatterjee

I design systems that teams actually build and run. I help growing businesses turn AI ideas into secure, practical systems that actually work — drawing on 20+ years designing and delivering enterprise data and AI platforms, from ground-up builds through to complex modernisation.

Let's talk about your AI & Data roadmap
  • $9MSavings delivered via modernisation
  • 20+Years in enterprise tech
  • 100+Legacy systems modernised
Sectors I've worked across
Financial Services & Banking Insurance Home Health Higher Education Professional Services Resources & Mining
Portrait of Sayan Chatterjee
Open to Advisory & Architecture roles

Is this you?

You might need me if...

"We don't know where to start with AI, and every vendor pitch sounds like the answer."

How I can help: Cut through the noise — assess what's realistic for your business, and build a practical roadmap prioritised by actual value, not hype.

"We've got an AI prototype that worked in the demo, but it's not ready for production."

How I can help: Review what's been built, identify the gaps in security, data quality and reliability, and set out what it actually takes to get it live safely.

"A vendor has sent us a proposal for a data or AI platform, and we want an unbiased, technical opinion before we sign."

How I can help: Review the proposal on its technical merits — architecture, cost assumptions, lock-in risk — and give you a straight assessment you can trust.

"We need senior data & AI architecture judgement, but we're not ready to hire a full-time architect."

How I can help: Work as a fractional or advisory technical partner — bringing enterprise-grade thinking to your team on the timeline and budget that fits a growing business.

How I think

A simple way to get from idea to something real.

  1. 1

    Start with the business need

    What outcome are you actually trying to achieve — in plain terms, not the technology.

  2. 2

    Understand the data behind it

    What information exists, where it lives, how good it is, and what's missing.

  3. 3

    Design the platform

    A system shaped around that need and that data — not a generic template.

  4. 4

    Build in governance

    Guardrails from day one, so the system is safe, compliant and trustworthy.

  5. 5

    Automate wherever possible

    Reduce manual effort so the system keeps working reliably without constant hand-holding.

  6. 6

    Get it into production

    Ship it, support it, and make sure the people who use it every day can actually run it.

Why hire me

What you get when I join your team.

01

A board-ready AI & data strategy

A target-state architecture, governance model and platform roadmap your executives will approve — and your engineers can actually build against.

02

GenAI & agentic platforms that ship

Working GenAI patterns, agentic AI architecture, self-service machine learning platforms and AI-ready lakehouses — not proofs of concept that stall after the demo.

03

Governance that accelerates delivery

Guardrails, clear decision-making roles and review boards that satisfy regulators and risk teams while letting engineering move faster, not slower.

04

Executive trust and delivery follow-through

Someone who secures funding, builds and leads delivery teams, and stays with a program through to production — not just the strategy deck.

By the numbers

20+Years across ANZ Bank, Australian Unity and other enterprise technology programmes
42%Avg. productivity improvement in a published GitHub Copilot study, co-authored with ANZ Bank's Chief Technology Officer — led to a 1,000+ engineer rollout
$9M+In savings delivered
100+Legacy systems modernised
$250K+In ongoing annual savings

Selected work

What I've built, not just designed.

Data & AI

Flagship Research

Proving GitHub Copilot's Impact at Enterprise Scale

ANZ Bank · 2021–2024 · Published research (arXiv:2402.05636)

Problem: ANZ Bank, with an engineering workforce of 5,000+, wanted rigorous, data-driven evidence — not just vendor claims — on whether an AI pair-programming tool actually improved developer productivity, code quality and security, before rolling it out broadly in a regulated banking environment with real IP and data-security concerns.

Approach: Designed and ran a six-week controlled experiment across 100+ engineers split into Control and Copilot groups, using A/B testing methodology — two weeks of baseline preparation followed by four weeks of active testing with weekly coding challenges. Measured results using GitHub activity metrics, participant surveys, SonarQube static analysis and manual grading, then analysed the data with rigorous non-parametric statistical testing (Wilcoxon Signed Rank Test at 95% confidence). Co-authored the findings into a formal research paper alongside ANZ Bank's Chief Technology Officer and other senior engineering leaders.

Outcome: Found a 42% average improvement in task completion time (rising to 52% for less experienced engineers), alongside statistically significant reductions in bugs and code smells. The findings directly informed ANZ's decision to move to a production rollout across 1,000+ engineers, and the work earned personal appreciation from ANZ Bank's Chief Technology Officer.

Read the published paper (arXiv:2402.05636) ↗
Applied AI & Machine Learning

Building a Recommendation Engine for a Social-Impact Marketplace

An online volunteering marketplace

Problem: An online marketplace connecting skilled volunteers to non-profits and charities relied on manual keyword search, which meant relevant opportunities were often invisible to the right people — around 24% of jobs posted by charities received no applications at all, hurting both volunteer engagement and the charities' ability to get help.

Approach: As part of a small project team, I led a structured, end-to-end data science process — from understanding the real business need with the company's Chief Executive Officer, through making sense of small, messy data spread across 64 unstructured files with no schema, to building and rigorously comparing multiple candidate recommendation techniques before settling on the one best suited to the data actually available, rather than the most fashionable one.

Outcome: Delivered a working, practical recommendation approach the business could put into action immediately, plus a clear roadmap for a more advanced approach to adopt once the platform's data volume grows — an evidence-based "what now, what next" answer rather than a one-size-fits-all recommendation.

Data Lifecycle Management, Cloud & Infrastructure Architecture

Greenfield · Data Platform Architecture

Laying the Foundation for Federated Data Ownership at ANZ Bank

ANZ Bank

Problem: ANZ's data platform and data ownership were fully centralised under a single team, with every change or access request routed through that one team — creating a bottleneck that slowed delivery for every domain that depended on it.

Approach: Owned the design and implementation of ANZ's data mesh architecture from the ground up, including a data product catalogue and a data mesh control plane, establishing federated data ownership so domain teams could publish and consume trusted data products rather than routing everything through a central bottleneck.

Outcome: Laid the architectural foundation for federated data ownership across the bank, establishing the foundation that enabled self-serve analytics at scale — foundational work that other teams and future data and AI initiatives could build on, not a single point solution.

Infrastructure & Cloud Architecture

Bringing Data-Driven Decisions to Asset Lifecycle Management

ANZ Bank

Problem: ANZ's technology estate lacked clear visibility into its assets — what existed, its condition, cost and remaining useful life — making it hard to make confident, well-timed renewal and investment decisions.

Approach: As the architect responsible for this area, established an operating framework for Asset Lifecycle Management, bringing structured, data-driven decision-making to a process that had previously relied on incomplete, fragmented information.

Outcome: Delivered substantial, measurable cost savings through improved asset transparency — turning infrastructure decisions from guesswork into a repeatable, evidence-based process.

Legacy Modernisation

Modernising ANZ Bank's Legacy Technology Footprint

ANZ Bank

Problem: A large number of business-critical applications were running on end-of-life infrastructure and were no longer aligned to ANZ's architecture standards, creating growing risk and technical debt across the estate.

Approach: Led a team of senior architects through a programme that was deliberately architecture-first rather than an infrastructure-focused lift-and-shift — modernising individual applications against current architecture standards, supported by a repeatable Migration-as-a-Service delivery model, and chairing the programme's design governance forums.

Outcome: Modernised a large portfolio of business-critical applications into alignment with architecture standards, enabling the underlying legacy infrastructure to be fully decommissioned.

Expertise

Where I create the most value.

◆

Enterprise AI Strategy & Architecture

Data & AI strategy, reference architectures, architecture principles and standards, guardrails, patterns, and enterprise architecture governance.

◆

AI, GenAI & Agentic Platforms

Machine learning platforms, Generative AI, agentic AI architecture, large language model-enabled engineering, intelligent automation, responsible AI controls and governance patterns.

◆

AI-Ready Data Foundations

Lakehouse & medallion architectures, data mesh, semantic layers, ontology, knowledge graphs, metadata intelligence, lineage, data quality, privacy-by-design.

◆

Enterprise Investment & Governance

AI portfolio visibility, technology rationalisation, architecture review boards, business cases, value realisation, request for proposal leadership and vendor evaluation.

◆

Cloud, Security & Delivery

AWS, GCP, Azure, multi-cloud data architecture, secure cloud patterns, identity and access management, encryption, streaming & batch integration, continuous integration and deployment, regulatory-aligned controls.

◆

Leadership & Executive Influence

Architecture leadership, team leadership, stakeholder influence, governance forums, capability uplift and senior technology advisory.

Recognition

Publications, certifications & awards.

Published Research

The Impact of AI Tool on Engineering at ANZ Bank: An Empirical Study on GitHub Copilot within Corporate Environment

Co-authored with ANZ Bank's Chief Technology Officer. A controlled study across 100+ engineers found a 42% average productivity improvement using GitHub Copilot — a result that led ANZ to roll the tool out to 1,000+ engineers, and earned personal appreciation and recognition from ANZ Bank's Chief Technology Officer.

Read the paper on arXiv (2402.05636) ↗

Other Publications & Research

  • Machine Learning Recommender System — Online Volunteering Marketplace Matching volunteers to charity jobs using collaborative & content-based filtering.
  • Hindi Character Recognition System Unsupervised classification of handwritten characters — Indian Statistical Institute, Kolkata.

Certifications

  • The Open Group Architecture Framework (TOGAF) 9 Certified
  • AWS Certified Cloud Practitioner
  • Microsoft Certified Solutions AssociateCloud Platform
  • Architecting & Developing Microsoft Azure Solutions

Awards

  • Design Leadership AwardInternet Banking (AU) Modernisation, ANZ Bank
  • Banking APAC Excellence AwardExcellence, 2017
  • Outstanding Performance Award2016 & 2014

About

The person behind the architecture.

I'm an Enterprise Data & AI Leader based in Melbourne, Australia. Over 20+ years I've moved from hands-on software engineering to shaping enterprise technology strategy — always staying close enough to delivery that my strategies survive contact with real engineering teams, real regulators, and real budgets.

I care about architecture that ships: platforms that get adopted, governance that speeds teams up instead of slowing them down, and AI investment that's aligned to outcomes an executive can defend to their board. That's what I bring to every organisation I work with.

Education

Master of Data Science — University of Melbourne (2021)

Bachelor of Engineering in Information Technology

Based in

Melbourne, Australia

Australian Citizen — available for local & remote engagements

Beyond the work

Life outside the architecture diagrams.

Contact

Have a data or AI problem worth solving?

Whether it's enterprise AI strategy, a lakehouse rebuild, agentic AI adoption, or an architecture review your organisation needs — I'd like to hear about it.

30-minute initial conversation — no obligation.

Engagements are scoped to the complexity and outcome required. After an initial conversation, I provide a clear proposal and fixed fee before any paid work begins.

Prefer email? sayan@sayanchatterjee.com

LinkedIn in/chatterjee-sayan Location Melbourne, Australia