Production AI · Payments platforms

I lead product and technology for production AI and payments platforms.

I decide what gets built, how it gets built and what it costs, then lead the team that delivers it. In AI, I drive what decides whether it holds up in production: testing, cost, guardrails and human review. In payments, I take a business from launch to scale and keep every transaction matched to the books.

Open to hands-on product roles in production AI and payments · Based in Dubai

Ex-Delivery Hero · TabSquare.AI, acquired by Delivery Hero

Delivery HeroTabSquare.AI
Production AI
2x
Leads turned into customers with an AI sales assistant
Prodaptix client outcome · I owned the product
Production AI
30–40%
Lower AI cost per conversation on the same assistant
Model routing and bounded context · I owned the product
Payments
$50M
Payments processed a month, six months after launch (from $0)
Delivery Hero company outcome · I led product strategy and launch
Payments
85%
Fewer payments that did not match the books
Delivery Hero company outcome · reconciliation built as a service

Why talk to me

Good reasons to start a conversation

  • You want AI that works in real life, not just in a demo.

    An AI sales assistant I owned doubled the share of leads that became customers, and I shipped an AI digital worker for refunds and chargebacks with a human approval step. Every release was tested, and a person reviewed its work. I owned the production side too: data, evals, MLOps trade-offs, cost, and governance with guardrails and a test harness.

  • You're launching or fixing payments.

    I took a payments business from $0 to $50M a month in 6 months and cut payments that didn't match the books by 85%. My work spans Tap to Pay, online payments and in-person terminals, with 25+ payment methods including alternative payments enabled. I also created the open API that let partners connect to the platform, which drove adoption.

  • You need product and engineering to move as one team.

    I led product inside a 100+ engineer organisation, working with CTOs and engineering leads. I led 10+ engineers directly at TabSquare AI and managed 6 product managers at Delivery Hero, so I can lead both the engineering and the product side. I managed the core platform, including the move from a monolith to microservices, the partner middleware, and developer experience and APIs.

I’m open to hands-on product roles in production AI and payments. Based in Dubai.

Talk about a role

Applied AI in production

Shipping an AI demo is easy. Running it in production is the job.

AI in production since 2017: recommendations, fraud detection, AI sales assistants and agents. These are the problems I own once it’s live.

  • 01

    Testing before release

    Every release has to pass 300 real example questions. They check accuracy, catch problems when business rules change, and test answers against company policy so the AI doesn't make things up.

    300-question test set

  • 02

    Cost

    Each task goes to the right-sized AI model: a smaller, cheaper one for summaries and after-call work, a stronger one for live calls and deal numbers. Limiting how much the AI reads per answer cut the cost further.

    30–40% lower cost per conversation

  • 03

    Safety limits

    AI agents can read data but never change it. Refunds and chargebacks need a person's approval, and a supervisor reviews the sales assistant's output.

    A person in the loop on money and customers

  • 04

    Accuracy

    The assistant only looks at the rep's own accounts, and any figure the AI pulls out is checked against the source before anyone relies on it.

    Open-source payments-ops-mcp

  • 05

    Memory

    The assistant remembers the current conversation and past account history, so follow-ups don't start from zero, while limiting what it reads each time.

    2x more leads converted

  • 06

    Data first

    I plan what to track before any AI model is trained, because AI is only as good as the data it learns from.

    In production since 2017 · +8% order value

See the Applied AI deep dive

Payments platforms

Launching payments is the start. Keeping them reliable, matched and profitable is the job.

Payments launched, run and reconciled across 10+ markets at Delivery Hero, then embedded payments for SaaS clients through Prodaptix. Results below are company outcomes; the text says what I owned.

  • 01

    Launch

    Launched a PayFac with engineering, risk and finance, starting in one narrow market to prove the economics and reliability before expanding. It went from about $10M a month in its first two months to $50M within six.

    $50M a month within 6 months, from $0

  • 02

    Acceptance

    Online checkout with saved cards and one-click payment, in-person terminals, Tap to Pay on iPhone and Android, and stored-value wallets.

    25+ payment methods

  • 03

    Routing

    Smart routing across payment providers, built so one slow provider cannot take down checkout.

    20+ payment providers

  • 04

    Reconciliation

    Every payment matched between the provider, the bank statement and the ledger. Treated as a service from day one, not a back-office report.

    85% fewer unreconciled items

  • 05

    Settlement & payouts

    Batch close, settlement and merchant payouts, with offline-first terminals that never double-charge when the network drops.

    70% faster settlement close

  • 06

    Payments P&L

    Co-owned the payments P&L across 10+ markets: buy and sell rates, pricing and staffing decisions.

    Contributed to 28% payments revenue growth

Where payments meets production AI

payments-ops-mcp

An open-source tool that lets an AI assistant investigate payment mismatches and disputes and recommend actions, but cannot move money. Payments discipline (idempotency, audit trails, least-privilege access, human approval) applied to AI agents. My own project.

Open source · 50 tests

Read the article
See the Payments deep dive

Platform foundation

The shared platform underneath both

Production AI and payments both ran on shared platform work: the multi-tenant platform behind 8,000 merchants at TabSquare AI, and open APIs and a developer portal at Delivery Hero. These are supporting capabilities, kept live for context.

Who owned what: ordering, loyalty, partner integrations and recommendations: TabSquare AI. Payments, routing, fraud scoring and the API platform: Delivery Hero, after it acquired TabSquare. RAG assistant and the refunds and chargebacks digital worker: Prodaptix client work. payments-ops-mcp: my open-source project.

Selected Work

Selected work: one case per pillar

The problem, what I owned, and what changed.

AI · Sales assistant

An AI assistant that helps sales reps close more deals

Prodaptix · Embedded with client · 2024–present

Problem

Reps lost the thread between calls, and AI answers were slow, costly and hard to trust.

What I owned

I owned the product: what the assistant remembers, what it is allowed to see, which AI model handles which task, and how every release is tested. A supervisor reviews its output.

Company outcome

2x more leads converted with an AI sales assistant

Read the full case

Payments · New business

Launching a payments business from zero

Delivery Hero · Regional Payments · 2021–2024

Problem

Many payment providers and fragile connections made payments hard to trust and hard to match to the books across a large marketplace.

What I owned

I owned the product strategy and launch plan with engineering, risk and finance: a reliable payments system that never double-charges and keeps every transaction matched to the books.

Company outcome

$50M

a month in payments, within six months of launch, from $0

Read the full case

How I decide

Where the trade-offs sit

Leading product and technology means making calls where good options pull in different directions. Here is one decision from each area, what it cost, and what it delivered.

Product

Team independence vs more to run

The call
Split one large system into smaller parts, each owned by a team, with a step-by-step migration, so regional teams could release on their own.
What it cost
More running costs and more complexity to manage.
What it delivered
Regional teams released independently and the platform could change one part at a time.
See the case

Business

Launch speed vs reliability

The call
Built the payments system so one slow provider could not take down checkout, and launched in one narrow market first to prove it made money before expanding.
What it cost
Some delay in matching payments to the books, in exchange for checkout that kept working when a provider was slow.
What it delivered
$50M a month in payments within 6 months, from about $10M a month in the first two. Payments that didn't match the books fell 85% and settlement closed 70% faster.
See the case

Data

Build now vs get the data right first

The call
Planned the data before building the AI: 25+ tracked events, with every suggestion labelled by where it appeared and what the customer did next.
What it delivered
+8% average order value across 8,000 merchants.
See the case

Technology

Bigger context vs bounded context

The call
Bounded what the assistant reads per answer by relevance, recency and task, instead of letting the prompt keep growing.
What it cost
Some risk of dropping relevant account history, so recall and grounding are tested alongside cost and speed.
What it delivered
30–40% lower cost per conversation on the sales assistant.
See the case

AI

Autonomy vs control, accuracy vs cost

The call
No AI ships without a test it must pass, a cost and speed budget, and monitoring. Agents can read data but never change it, and money moves only with a person's approval.
What it cost
Some raw accuracy, to keep the AI fast and affordable.
What it delivered
2x more leads converted and 30–40% lower cost per conversation on the sales assistant.
See the case

The architecture log

10 more decisions, each with the problem, what I chose, the trade-off and who I worked with, including CTOs and engineering managers.

Read the log

What colleagues say

From people who worked with me on payments, product and platform work.

Payments migration
I have worked with Janahan quite extensively and on one notable project, he was instrumental in ensuring that the transition from one payments provider to another, was timely and smooth. That was an extremely complex task involving a new product, multiple stakeholders, and sensitive and aggressive deadlines. In the first 2 months of the launch itself, the product reached ~$10M in volumes!

Kanika Mittal

VP Fintech & Partner, Travel Payments

Product intuition
What stood out consistently was his sharp product intuition and his ability to craft solutions that were not just functional, but genuinely aligned to market needs. He has a rare ability to listen deeply, to customers, and to the market and translate that into extraordinary product outcomes. He was a true partner in every sense, collaborative, grounded in data, and always thinking a few steps ahead.

Sameer Kaushik

VP Sales, TabSquare AI

Engineering + product
What makes Janahan stand out is his unique combination of technical depth and product vision. With his engineering background, he not only understands what it takes to build stable and scalable products but also brings great ideas that are practical, intuitive, and user-friendly.

Kar How Wong

Regional VP Operations

More recommendations on LinkedIn →

Where I've worked

Worked at

Prodaptix
Delivery Hero
TabSquare.AI
Aigens
42Race
X-ONT Software
Visni
David Pieris Motor Company

Let's talk

Open to hands-on product roles in production AI and payments · Based in Dubai