The short version

We embed staff-level engineers into client teams at banks, fintechs, energy majors and government across APAC. We're deepening the data and applied AI side of that, and we need people senior enough to be the only data person in a room full of client stakeholders and still be the most useful person there.

This is not a research role. Models that never leave a notebook don't count here. Neither do decks.

— 01 · What you'd actually do

- Own the end-to-end arc on a client engagement: design, delivery, and the governance that lets it survive an audit.
- Sit inside a client's workplace — their lakehouse, their repo, their standups — and work to their standards rather than importing your own.
- Build production data pipelines and the analytics on top of them. Most engagements need someone who can fix the pipeline before modelling anything.
- Ship GenAI systems that hold up: retrieval that returns the right document, evals that catch regressions, guardrails and observability so a client can tell whether their AI feature is improving or just getting louder.
- Take a vague commercial problem — "our claims triage is too slow", "we can't forecast demand at SKU level", "the traders don't trust the model" — and turn it into something specific enough to build, or say plainly when modelling isn't the answer.
- Write the governance narrative. Our clients get audited. Model cards, lineage, decision logs and the reasoning behind a feature choice are part of the deliverable, not overhead.
- Explain the same system to a risk officer, a plant engineer and a CTO in one week, adjusting for each without dumbing it down.

— 02 · What we need
- 5+ years building data and AI systems, with a meaningful share of that in production rather than analysis.
- Databricks depth. We're in the Databricks partner program and it's central to how we deliver — Delta, Unity Catalog, MLflow, workflow orchestration, and the cost conversations that come with all of it. Equivalent depth in another lakehouse platform, with willingness to convert, is fine.




- Strong Python and SQL. Comfortable in someone else's codebase.
- Real depth in at least one of: forecasting and demand planning, risk and actuarial modelling, fraud and anomaly detection, or applied LLM systems.
- Industry exposure to energy, insurance, financial services or FMCG. These are the sectors where our work lands, and the domain vocabulary matters more than people expect.
- The judgement to reach for logistic regression when logistic regression is the right answer.
- Client-facing composure. You'll be in front of stakeholders without an account manager translating for you.
- AI-fluent in your own work. We're an Anthropic technology partner and use Claude Code, Cursor and Copilot daily. We're interested in how you use them, where you deliberately don't, and how you check the output.

— 03 · Nice to have
- SAP or other enterprise-system interoperability work — the integration layer is usually where these projects stall.
- Postgraduate study in data science, ML or a quantitative discipline.
- Professional membership (BCS, ACS) or vendor accreditations you actually keep current.
- Experience being the first data hire on an engagement, with no existing platform to inherit.

— 04 · The Lead delta Everything above, plus:
- Technical ownership of a pod of 3–6 across a full engagement, including the Colombo engineering team.
- Scoping and estimation with commercial consequences — you'll shape what we commit to, not just deliver it.
- Being the architect of record: choosing the platform pattern, defending it to a client's enterprise architecture group, and living with the consequences.
- Contributing to how we sell this — reference architectures, evaluation playbooks, the patterns that make the next engagement faster.

— 05 · Process
1. Intro call — 30 minutes, mutual.
2. Technical conversation — a real problem from client work, discussed rather than performed. No take-home.
3. Client-scenario session — you present an architecture and defend it under questioning, the way you would in week two.
4. Conversation with a partner.

Compensation is discussed openly at the first call.

📌 Senior / Lead Data Scientist — AI & Data Engineering (Full Time / Fractional) (Kolamba)
🏢 Rhino Partners
📍 Kolamba