L Lasith Jayarathne
ExperienceAI BuildsWritingAboutContact Advisory →
London — GMT+0 2026
PRINCIPAL PRODUCT MANAGER
PMP · SAFe POPM · AIGP IN PROGRESS · LONDON · GMT+0
lasithj1@gmail.com
13+ Years setting strategy
and delivery
Domains Aviation · Enterprise · Public sector
Certifications PMP · SAFe POPM · AIGP in progress
The trio Build · Write · Lead

Principal Product Manager

AI Products · Enterprise Delivery
I bring AI fluency, product vision and enterprise leadership to the same role.
Thirteen years setting strategy and leading global delivery across aviation, enterprise platforms and public sector. I build AI products to keep the thinking sharp. I write to pressure-test it in public. The three make each other sharper.
Scroll
§ 01 Experience & Capabilities
Eight surfaces · scan and expand what's relevant

Eight domains. Scan the headers, expand what's relevant.

Defining where AI products need to go — translating AI capability into clear product vision, multi-year strategy and roadmaps that create genuine business value. Identifying where AI fundamentally changes a product, where it enhances it, and where human judgement must remain in the loop. Setting a direction that engineering, design and senior stakeholders can align behind and execute against with confidence.

Designing AI products from the ground up — shaping the user experience, evaluation criteria, confidence thresholds and model performance requirements as product decisions, not engineering ones. Defining what good looks like for AI features before a line of code is written, and maintaining that standard through delivery, launch and iteration.

Building AI governance into the product from the start — not as a compliance layer added at the end, but as a design discipline that makes AI systems more trustworthy, more defensible and more durable. Working knowledge of NIST AI Risk Management Framework, ISO 42001 and EU AI Act obligations. Practical application of risk-tiered AI classification, human-in-the-loop design, explainability requirements, fairness considerations and auditability as first-class product requirements. AIGP certification in progress.

Applying AI-native workflows across the full product lifecycle — discovery, UX, requirements, prototyping and delivery. Using modern AI tooling and virtual AI pods to compress the cycle from ambiguous problem to testable product. Hands-on builder of AI systems, keeping the strategy grounded in what is actually possible and current.

Leading product strategy and global delivery across complex, multi-team organisations. Setting roadmaps, defining OKRs, making investment decisions and aligning senior stakeholders across multiple functions and geographies. Operating at portfolio scale with competing priorities, long time horizons and high organisational complexity — and maintaining clarity of direction throughout.

Managing large-scale product portfolios spanning multiple teams and geographies. Operating model design, lean budgeting, dependency management, PI planning and delivery governance across globally distributed engineering organisations. Balancing product evolution, technical debt, customer delivery and platform sustainability across a complex, multi-year delivery programme.

Deep domain knowledge across airline distribution, operations, ancillary services and enterprise aviation platforms. Product definition, systems integration and delivery across major airline and travel technology organisations. Strong understanding of the commercial, operational and regulatory constraints that shape product decisions in the aviation industry.

Product strategy, vision and delivery for national-scale government platforms — identity systems, citizen services and revenue platforms. Navigating complex multi-stakeholder environments spanning government agencies, technology partners and service providers. Designing products where trust, security and accessibility are non-negotiable requirements from day one.

§ 02 AI Builds
Personal projects · continuous learning · two categories

Built to stay at the frontier of what's possible.

Four projects across two categories — AI Product Case Studies where AI is the core design problem, and AI-Enabled Builds where AI accelerates the work.

Category 01

AI Product Case Studies

Products where AI is the core design problem — the architecture, the decisions and the governance are all AI-specific.

BUILD · 01 NexBridge
Open-source AI-governed middleware framework

A framework for modernising legacy enterprise integrations behind a risk-tiered classification system — T1 through T4 — where each tier determines the level of AI autonomy, human oversight and audit required. Dual-agent verification, human escalation gates, confidence thresholds and immutable audit logs built in by design. Built on LangGraph, FastAPI and the Anthropic Claude API. Demonstrates AI governance designed as a product architecture decision from the start — not retrofitted as compliance.

LangGraphFastAPIAnthropic Claude APIOllamaLlamaPythonRisk-tiered classificationHuman-in-the-loop designMulti-agent architectureAudit logging
BUILD · 02 Multi-Agent AI Investor System
Explainable multi-agent decisioning prototype

Six specialised agents — technical, fundamental, sentiment, seasonality, competitor and decision — working in a layered orchestration model to produce explainable investment signals. Confidence scoring, decision drivers, caution flags and visible data gaps are first-class outputs, not buried in logs. Designed around user trust and explainability as core product requirements from day one. Demonstrates multi-agent system design, orchestration architecture and the product discipline of making AI reasoning visible to the user.

Multi-agent orchestrationLLM integrationConfidence scoringExplainability designSignal aggregationRAGPrompt engineeringAgent supervision
Category 02

AI-Enabled Builds

Products built using AI as a delivery accelerator — AI-native workflows applied to real product problems.

BUILD · 03 Internal SAFe Train Management Tool
AI-enabled product operations build

A product operations tool built using AI-native workflows — Claude, Windsurf and virtual AI pods — to improve planning, dependency and delivery visibility across a large-scale Agile release train. Explored how AI-native delivery compresses the cycle from problem to working product. Demonstrates the practical application of AI-native ways of working to unglamorous but high-value operational problems.

ClaudeWindsurfVirtual AI podsAI-native workflowsPrompt engineeringAI-assisted requirementsRapid prototyping
BUILD · 04 Crikly
AI-assisted zero-to-one marketplace concept

End-to-end product thinking for a two-sided coaching marketplace — user problems, role-based journeys, MVP scope, booking flows, onboarding and go-to-market positioning — developed using AI-assisted discovery, journey mapping and prototyping. Taken from ambiguous problem space to a fully defined, buildable product concept. Demonstrates zero-to-one product thinking and marketplace design applied outside the day job, with AI accelerating the discovery cycle.

AI-assisted discoveryJourney mappingMVP scopingAI-accelerated prototypingGo-to-market design
§ 03 Writing
Responsible by Design — monthly · all articles link out

Where the thinking gets pressure-tested in public.

Responsible by Design — a monthly newsletter on AI product strategy and responsible AI. All articles link out to LinkedIn in a new tab. Newest first.

§ 04 About

Aviation. Enterprise. Public sector.

I started product-side in aviation — airline distribution systems, ancillary services, operational tooling — where the cost of a missed detail is measured in flights and people, not story points. That grounding stays with you.

From there, enterprise platforms and public sector digital work taught me that the unglamorous parts of a product — the intake, the escalation paths, the audit trail — are where trust is actually built or lost. The governance thinking didn’t come from frameworks. It came from building products where the consequences of getting it wrong were real.

Today I work at the intersection of AI product strategy and enterprise delivery. I build AI systems personally to keep the thinking current, write publicly to pressure-test the ideas, and lead global product teams to ship. The three disciplines inform each other — and that combination is what I bring to every role.

Education
  • BSc Computer Science Rajarata University of Sri Lanka · 2007 — 2011
Certifications
  • PMP — Project Management Professional PMI
  • SAFe® Product Owner / Product Manager Scaled Agile
  • AIGP — AI Governance Professional In progress IAPP
§ 05 Contact
Senior product leadership conversations

Let's talk.

Open to senior product leadership conversations — Principal and Lead AI PM roles, and enterprise product leadership where AI strategy is central to the work.

I'm based in London and require UK visa sponsorship. I mention it early so the conversation starts on honest ground.

Open toPrincipal · Lead AI PM
BasedLondon, UK
VisaRequires UK visa sponsorship
CertificationsPMP · SAFe POPM · AIGP (in progress)
Response≤ 48 hours