ML Architecture Review

How I Can Help

Advisory Call

$600 · 90 minutes - book call

A focused conversation about a specific challenge you’re facing. You bring the context, I bring experience from building ML systems at scale.

Good for:

  • Sanity-checking a technical decision

  • “Should we build or buy” questions

  • Getting unstuck on a specific problem

  • Exploring whether a deeper engagement makes sense

What you get: 90-minute video call + summary notes with key recommendations.

Focused Review

$4,000 · 1 week - book intro call

I diagnose the problem and tell you what to fix, fast.

Good for:

  • “Our serving latency is killing us”

  • “Is our training pipeline set up correctly?”

  • “We’re about to launch – what are we missing?”

  • Pre-investment technical due diligence

What you get:

  • 60-min scoping call to understand the problem

  • Async review of your architecture, code, and docs

  • Action list: 3-5 prioritized fixes with implementation notes

  • 30-min debrief to answer questions and clarify next steps


Comprehensive Audit

$12,000 · 2-3 weeks - book intro call

I assess your entire ML stack and build a roadmap to fix it.

Good for:

  • Platform-wide health check before major investment

  • Leadership wanting an outside perspective on systemic issues

  • Teams inheriting legacy ML systems

  • Post-mortem after a failed ML initiative

What you get:

  • Everything in the Focused Review, but across multiple systems

  • Stakeholder interviews with engineering, product, and leadership

  • Formal written report: technical findings, organizational gaps, and risk assessment

  • 6-month roadmap with effort estimates, dependencies, and sequencing

  • Follow-up call 4 weeks later to check progress

What I Look At

Every review is tailored to your situation, but I typically assess:

Architecture & Design – Model serving setup, training infrastructure, feature pipelines, online/offline consistency, experimentation systems.

Operational Health – Monitoring and alerting, model freshness, debugging and rollback capabilities, incident response readiness.

Organizational Patterns – Team structure, ownership models, tech debt tradeoffs, build vs buy decisions.

How It Works

  1. Reach out – Use the links above or reply to this page with a brief description of your situation.

  2. Quick chat – We’ll have a 15-minute call to see if I can help and scope the engagement.

  3. We go – Once we agree on scope, I’ll send an invoice and we’ll get started.


Questions?

What if I’m not sure which tier I need? Start with an advisory call. We’ll figure out the right scope together.

Do you sign NDAs? Yes, happy to sign a mutual NDA before we discuss anything sensitive.

What’s your availability? I take on 2-3 reviews per month alongside my full-time role. Reach out early if you have a timeline in mind.

Can you help with implementation after the review? My focus is on the review and recommendations. I can point you in the right direction for implementation support if needed.