Principal Product Designer & Builder

I design AI-nativeproducts & systems people & AI agents.

15+ years· Capital One · Google · Walmart | End-to-End Product Design | Shipped AI / ML products since 2018
Agentic systems· commerce · developer tools · complex enterprise B2B platforms and mobile apps
Design ↔ code· Design systems · React · APIs · GitHub · AI evaluations and release workflows
GitHub · Last 30 days318+ contributions
Featured work
Capital One Software2026

Context-Aware Agentic Investigations

Agentic AI0→1End-to-End UXEnterprise B2BCloud data platform
My questionHow can AI help detect cost spikes and act early?

I designed an agentic investigation experience for cloud data teams, data engineers, building trust through evidence, confidence signals, human oversight, and validation.

ContextLogs & telemetry · cost/performance signals · lineage · user intent
Agent reasoningHypotheses · next action · supporting evidence
Human controlReview · redirect · verify · recover
OutcomeRoot cause, explanation and next best action

25–40% faster root-cause and optimization workflows 4–5 tools consolidated into one investigation experience Shipped to Private Preview

Slingshot enterprise AI investigation interface
Independent AI products & systems

Self-directed work where I explore emerging AI interaction models, build the system, and test the product behavior.

Agentic Commerce - Autonomous Buying

Buyer’s AgentWorking Sandbox demo0 → 1

Repeated household purchases led me to explore emerging commerce agents and a memory-first buyer agent.

My questionHow much purchasing authority should an AI agent actually have?
What I exploredShopify agentic commerce, Muse connector eligibility, UCP, ACP, Google AP2, Visa TAP, and direct retailer APIs.
Designed, built & verifiedA successful controlled autonomous purchase: WooCommerce demo store & Stripe.A cart-ready checkout experience (no payment executed): Shopify UCP.Idea → working prototype in ~3 weeks
💡Production autonomous purchasing requires retailer and payment access.
Buyer's Agent mobile interface
Buyer's Agent purchase decision interface
Buyer's Agent product decision detail

How Buyer’s Agent works

RememberHousehold preferences + confirmed purchase history
Authorize
AUTHORIZED · ASK · BLOCK
ExecuteBuy within rules · Verify · Record

Conversation Design for Agentic Systems

Conversation DesignOpen-sourceAI Native UX

I designed conversation behavior and human–agent interaction for real problems in agentic systems—not chat UI design. I translated those problems into conversation-state patterns across open-source AI products, including merged and released fix.

Key themes
StateAuthorityApprovalTurn-takingRecoveryTrust
Contributed to
OpenAI Codex
assistant-ui
CopilotKit
Prototyped in
Voiceflow

My opensource contributions

5+Open-source interaction issues
3+AI products contributed to
1+Merged + released fix
Read case study →See on GitHub ↗
A
Agent

I’m ready to optimize the warehouse. This will update 3 tables.

Approval required

Run optimization on 3 tables?

Cancelled

Conversation moved on. This approval is no longer valid.

Y
You

Actually, let’s check one more thing first.

A
Agent

Got it. I’ll check the latest data and compare it with last week.

A
Agent

Here’s the comparison. Want me to suggest an updated plan?

Y
You

Yes. Update the plan with the new comparison.

A
Agent

Updated. The previous approval stays cancelled. I’ll ask again before anything runs.

assistant-ui · merged + releasedConversation state: Approval synchronized with active thread.

Agent-Ready Design Systems

Design SystemsAI-ReadyProductCodeGovernance

A practical, end-to-end design system for humans and AI agents — built, applied, and operationalized across design, code, and agent workflows.

BuildCreated a public React design system with tokens, components, contracts, registry, governance, and an agent harness.
ApplyApplied the same system thinking through an external open-source design-system contribution.
OperationalizeConnected Figma ↔ code ↔ agents with variables, component mappings, MCP/context, validation, and drift checks.
Read case studyRead the playbook
Portfolio Design System
FigmaDesign
  • Variables
  • Components
  • Guidelines
ReactCode
  • Tokens
  • Components
  • Tests
Agent harnessAI
  • MCP + skills
  • Context
  • Validation
Agent workflow
Plan→Discover→Reuse / Extend→Validate→Ship
Evaluation harnessControlled tasks to measure agent use of the design system.
Task setWith vs without contextReuse + accuracyA11y checksLogs + traces
[N]shared components[N]versioned releases[X%]component reuse[X/N]agent task success
Interested in roles across AI-native products, agentic systems, developer tools, enterprise platforms and high-trust product experiences.

I'm always up for a good conversation.

Let’s talk →

Previous portfolios: 2000 · 2001 · 2024 · 2025

2026 © Kirubha Kittusamy