thaman✳︎
Say hello
THAMAN MUTHAPPA / WORK

The decisions behind the products.

8 case studies, written the way I write specs: the problem and its evidence, what was out of scope, the options I turned down, what shipped and what actually changed.

Quarterly close · Portfolio Engine
Weeks → days
Document review · Verification
Weeks → hours
Scenarios run, zero ledger writes
1,000+
Daily active users · Thal.E
1,000+
HOW I BUILD AI PRODUCTS

Three rules I keep
coming back to.

  1. Provenance is a requirement.

    If people need to trust an AI-assisted number, the evidence sits next to the number, not in an audit log.

    See it in Portfolio Engine ↗︎

  2. Draw the boundary before the interface.

    Decide what an AI interaction is allowed to change before deciding what it can say.

    See it in Conversational exit modeling ↗︎

  3. Stop work with evidence.

    Pausing a product is a decision with the same bar as shipping one.

    See it in Documents over Tables ↗︎
Jul 2025 – now

VectorShift

Agentic AI for private equity. Built as an engineer behind the pivot into financial services, then owned as the company’s first PM from April 2026.

SYSTEMS OF RECORDBuilt as engineer

Portfolio Engine

Make provenance a product requirement. A number cannot exist in the record without the document it came from.

Weeks → daysQuarterly close
AGENTIC REASONINGBuilt as engineer

Deal Intelligence

Break the thesis into linked claims, risks and initiatives, and design an agent that reasons across every deal the fund holds.

ThousandsInitiatives proposed and amended
CONVERSATIONAL AIBuilt as engineer

Conversational exit modeling

Let them ask in plain language, and keep every scenario architecturally separate from the ledger.

1,000+Scenarios run
AI VERIFICATION0 → 1

Verification

Check each claim and its citation as a pair, and give reviewers a list of issues instead of a document to reread.

Weeks → hoursReview workflow
ENTERPRISE PRODUCTPlatform foundation

Permissions that fit the work

Rebuild permissions as role-based access control mapped to the 5 personas who actually use the platform.

2 → 5Generic roles → persona roles
PRIORITISATIONDecision

Documents over Tables

Prioritise Documents and deprioritise Tables, based on usage data.

1 callDocuments in, Tables paused
Dec 2023 – Jul 2025

Infra.Market

Retail operations and sales tooling. The first full product loop, before the title.

DISCOVERY → DELIVERYEngineer, pre-title

Photon Studio

Start with store-manager interviews and stay with the problem through the spec, the launch and the feedback.

4 stagesInterview → spec → launch → feedback
May – Dec 2023

Thal.E

Consumer 0 → 1 across three campus canteens, as founding product lead.

CONSUMER 0 → 1Founding product lead

Thal.E

Build ordering and queue management for 3 canteens at Christ University, and own it from research to post-launch metrics.

1,000+Daily active users
SMALLER STORIES

Also shipped.

Real products and real numbers, without a full write-up.