Let's stabilise and scale your machine learning systems.

Fastest path: LinkedIn message with a few lines about your product, current stack, and where reliability or cost is hurting.

Message on LinkedIn Email me

[email protected]Download vCard

LinkedIn is fastest. Replies within 48 hours. Include timeline and where reliability, latency, or cost is hurting to speed triage.

Or send a message

Reply within 48 hours. Prefer a call? Include times.

Great Fit

Teams moving ML from prototype to reliable production
Teams needing architecture review and SLO-driven deployment planning
LLM teams fixing latency, cost, or reliability issues
Advisory-only support for early to growth-stage ML products

Engagement Models

Advisory blocks

Targeted reviews, incident postmortems, and roadmap validation for fast clarity.

Best for: 1–2 week fixed scope.

Stabilisation sprint

Triage reliability issues, lock in SLOs, and remove immediate cost or latency risk.

Best for: 4–8 weeks, outcome-focused.

Enablement + handoff

Pair with your team on deployments, guardrails, and handoff so reliability continues.

Best for: 8–16 weeks, includes coaching.

What to Include

Product and user context
Current stack and reliability/cost pain points
Clear success metric and timeline

Decision Checklist

SLOs and error budgets defined and agreed
Rollback paths validated in staging
Latency and unit-cost targets per user journey
Observability stories for “what broke” and “why”
Runbooks and on-call expectations written down

What Happens Next

1) Quick triage

I respond within 48 hours to confirm fit and next steps.

2) Deep dive

30-45 minute call to map goals, risks, and success metrics.

3) Plan

You get a concise plan with scope, timelines, and immediate wins.

Common Questions

How soon can we start?

For the right fit, I can usually begin discovery within 5 business days. Implementation start depends on scope, but I prioritise stabilising risk areas quickly.

Do you sign NDAs and work with regulated data?

Yes. I regularly work under mutual NDAs and can align to your data-handling rules. I avoid holding production data longer than necessary and prefer scoped access.

What engagement lengths do you support?

Typical engagements run 6–16 weeks: assess → stabilise → optimise → hand off. Shorter advisory blocks are possible for architecture reviews or incident postmortems.

How do you price?

Fixed-scope for well-defined deliverables; weekly for discovery/enablement; or milestone-based for outcome-driven work. Pricing is aligned to measurable reliability and latency targets.

What time zones do you work with?

I collaborate asynchronously with clear updates. Core overlap of 3–4 hours with US and EU teams is standard; APAC is fully covered.

Ready to ship ML systems that stay up?

If your team is serious about production reliability, latency targets, and cost discipline—let's talk.

Message on LinkedIn Email: [email protected]