From strategy and architecture to shipped product, working with founders and enterprises navigating the shift to agentic systems.
10+ years building · 40+ clients · Colombo, serving Asia & beyond
AVAILABLE · Colombo
UTC+5:30
Suresh Michael Peiris
AI Consultant & Architect
10+
Years building software
40+
Clients served
50+
Projects shipped
3
Continents worked across
What I do
Most engagements start with one of these and grow into the next. All of them end with your team owning the result.
Roadmaps, opportunity mapping, and build-vs-buy decisions for teams at the start of their AI journey.
Agentic pipelines, RAG systems, and LLM integration patterns designed to scale from day one.
Workshops and tooling that turn your engineers and product teams into effective AI practitioners.
Hands-on implementation, from prototype to production, working alongside your team.
How I work
01
A short, focused look at your workflows, data, and constraints. We end with a written view of what's actually feasible.
02
System design on paper before code: data flow, models, evaluation, failure modes, and what it will cost to run.
03
A thin, working slice in front of real users within weeks, enough to prove the hard parts, not just the demo path.
04
Hardening the parts that break: retrieval quality, guardrails, observability, and the cost curve at real volume.
05
Your team takes it from here. Documentation, walkthroughs, and the review habits that keep quality from drifting.
My approach
I won't promise what AI can't deliver. We start with a clear picture of feasibility, cost, and trade-offs.
The model is rarely the hard part. Data, latency, safety, and integration are where projects succeed or fail.
I bias toward working software over long strategy decks. We get something in front of users fast, then improve.
50+
Projects shipped
About
I'm Suresh, an AI consultant and architect helping companies move from AI curiosity to AI capability. Based in Colombo, Sri Lanka, I work with founders and enterprises across Asia and globally.
A decade of software engineering sits behind that, now pointed entirely at practical AI: strategy that doesn't oversell, architecture that holds under real load, and code that ships.
Workshops
Hands-on workshops and talks on voice agents, agentic systems, and LLM engineering, with slides from each session included.
A workshop on the full path from an empty Google Cloud project to a running service: picking the right compute product, wiring the build, deploying without downtime, and knowing what broke when something does.
August 20, 2026
A beginner's session on CTF competitions and the cloud security ideas underneath them: how the challenge categories work, the method for attacking one, and the misconfigurations that show up in real accounts as often as they do in contests.
August 20, 2026
A session on the part of competitive programming that isn't algorithms: how to read a problem, how to solve it on paper before you touch the keyboard, and what to practise between contests.
August 6, 2026
Writing
Practical pieces on AI strategy, architecture, and what actually survives contact with production.
Architectural notes on WebMCP: what the browser is actually offering, why the tool runs inside your user's session, and what that quietly asks you to accept.
August 21, 2026
Architectural notes on Microsoft's Agent Governance Toolkit: why a deterministic gate beats a well-worded prompt, and the composition gap nobody has solved.
August 20, 2026
Architectural notes from building a multi-tenant, AI-driven call centre platform: SIP interconnect, address governance, and tenant isolation.
August 5, 2026
Let's talk
Whether you're mapping out your first AI project or scaling a system that already ships, I'm happy to talk it through. No deck required.
Or email hello@sureshmichael.com