I don't advise on software. I ship it.

Ten years leading product at Orange, Fiserv, BNP Paribas and Aragon — now building and operating the systems myself, directing an AI engineering team. The three most recent are in production, and real money moves through two of them.

See the work How I work
Scroll

EXEGOVBUILDS SYSTEMSTHAT RUN

Most consulting ends at the deck. Mine ends at a deploy — automated trading on funded accounts, a CRM a field crew opens every morning, campaigns that launch themselves when a storm crosses a postcode.

The difference matters because software only reveals its real problems once it is carrying weight. A broker's API that under-reports a position, an order that silently refuses to cancel, a margin figure the vendor documents but never sends — none of that shows up in a workshop.

EXEGOVSHIPS WITH ANAI TEAM

I direct a team of AI engineering agents the way I used to direct people: I own what gets built and why, they own how, and every change is read by a human before it moves.

Design, implementation and adversarial review run in parallel — one agent builds, another tries to break the result. It is the reason a single person can carry three production systems, and the reason the work still gets reviewed like a team's.

EXEGOVPROVES IT WITHEVIDENCE

Every fix starts with a test confirmed failing, and every claim traces to a production log quoted rather than paraphrased. When something cannot be proven, it gets said plainly instead of dressed up.

That discipline is what let me show a broker that their own API returned two different values for one position fifteen seconds apart — with timestamps their support could not dismiss. Anything touching money stays behind a human gate.

What I take on

Three ways this usually starts.

All three end the same way: something running in production that somebody depends on, and a person who knows how to keep it there.

Build and operate

An idea, a spreadsheet, or a process that only works because one person remembers it. I take it to a deployed system — and then keep operating it: releases, monitoring, the unglamorous parts nobody quotes for.

Typically 6–16 weeks to first production
Rescue a stalled build

Six months in, no deploy, and everyone is still polite in the status meeting. I read the code as it actually is, name what is really blocking it, and get a working version in front of users before we argue about the rest.

Starts with a two-week read
Direct the AI team

For teams adopting AI engineering and finding that speed arrived without correctness. I set the gates: what gets specified, what gets tested, what a human must read before it ships, and what never runs unattended.

Fractional, part-time

Terms. Fixed price per feature when the scope is clear; hourly when it genuinely is not. Your repository, your cloud accounts, your keys — the work is yours from the first commit, and nothing about it depends on me staying.

Background

Ten years of owning what gets built.

Regulated payments, identity, telecom and open-source infrastructure, most of it in teams where a bad release had consequences. Today I run three lines of work at the same time: two for clients and one of my own.

Now Since 2026

Magdan Construction

Technology and marketing lead, contract 2026 to now · Chicago, IL

I run technology and marketing end to end for a Chicago disaster-restoration company. The owner decides what the business needs. I decide how it gets built, ship it, and keep it running.

  • CRM and ERP in one platform: lead intake, sales pipeline, project boards, insurance claims and inventory, plus finance, HR and e-signature modules. It now holds more than 200 clients and has taken over coordination work the office used to do by hand.
  • Native Android app for field crews: photos taken on site upload on their own into the right project and phase, and the crew can update the client record from the job.
  • AI layer: Claude-written client summaries over 2,800+ ingested emails, checked against a golden set before every release. It saves the office about 3 hours a week.
  • Marketing: Google Ads, including campaigns that switch on when the National Weather Service issues a storm alert for the service area; a new 39-page website with call tracking and a source on every lead; TCPA-compliant SMS; weather-triggered direct mail.
Next.jsPrismaPostgreSQLAndroid (Kotlin)Claude APIGoogle Ads APITwilioVercel

Algorithmic trading platform

Product lead and operator, client 2026 to now

A futures trading platform I built and operate for a private client. Bots trade funded accounts around the clock, and three people watch their own money on the dashboard.

  • About $1M of capital traded live 24/7 across 22 CME and ICE futures plus crypto perpetuals, through the Ironbeam and Coinbase APIs.
  • Layered risk controls. The decisions that carry real risk, like going live or changing limits, stay with a human.
  • An evaluation loop I designed: Claude agents review each day of trading and propose strategy and prompt changes, and a person approves every change before release.
  • Backtesting before release, 2,000+ automated tests and a CI-enforced 80% coverage gate.
PythonFastAPINext.jsREST + WebSocketIronbeamCoinbaseClaude agents

Exegov and StormFunnel

Founder, own products 2026 to now · Austin, TX

Two products of my own, built with a team of AI coding agents that I direct the way I used to direct engineering teams: I own what gets built and why, every change is read before it ships.

  • StormFunnel: a SaaS that turns National Weather Service alerts into ready Google Ads campaigns for home-services contractors. It reads alerts for all 3,143 US counties every 5 minutes, matches them to each customer's service area, and holds every campaign for approval before money moves. Specced and shipped solo since July 2026: 42 data models, 10,000+ automated tests, Stripe billing with prepaid ad wallets.
  • Exegov: one AI that a whole company shares, for owner-led businesses of 1 to 50 people. It knows the company's decisions and shared files, points out when a request goes against a rule someone has set, and keeps each person's own chats private. In build, ahead of launch.
Multi-tenant SaaSNext.jsGoogle Ads APIStripeNWS alertsClaude API
Before 2014 to 2025

Exegov

Founder and Lead Product Manager 2024 to 2025 · Austin, TX

Founded and led product for an agentic AI platform that gave LLM agents the business context they need to do real work inside a company. Took it from a blank page to paying customers.

  • Owned the product end to end: discovery with early-stage founders, problem framing, roadmap, pricing and launch.
  • Designed the core: a context layer that turned a company's documents, goals and data into structured context for AI agents (RAG over vector embeddings), so answers were grounded in that business, not generic.
  • Shipped multi-agent workflows in mid-2024, before agentic AI went mainstream: co-pilot agents for strategy, product, marketing, HR, finance and legal, orchestrated with AutoGen Studio, LangChain and Microsoft Prompt Flow.
  • Treated model choice as a product decision: benchmarked GPT-4, Gemma and DeepSeek on cost against answer quality per task, with prompt-level safeguards against hallucinations.
  • Led a six-person remote team across engineering, design and marketing.

Aragon

Lead Product Manager 2022 to 2024 · Remote

Open-source governance infrastructure used by DAOs including Lido, Curve and Decentraland. The role continued through the Aragon Association's 2023 transition into Aragon X.

  • Shipped Aragon OSx, a modular DAO framework, and Aragon App across Polygon, Arbitrum, Base and ZKsync: core governance infrastructure for thousands of on-chain organizations.
  • Led product delivery across about 70 contributors, a 30+ person core team plus community guilds, spanning protocol, frontend, UX and governance. Fully remote, async-first.
  • Merged two large crypto products into one roadmap and execution model, aligning smart-contract, frontend, ops and governance teams through shared milestones and OKRs.
  • Ran continuous experiments (A/B and quasi-experiments) on proposal flows, voter engagement and treasury UX: governance participation up 3.1x and proposal cycle time down 58% in 12 months.

BNP Paribas

Product Operations Lead 2019 to 2022 · Warsaw

Led the product transformation of a portfolio of 20+ products, with C-level sponsors and a direct team of seven.

  • Designed the adoption strategy and a custom target operating model: Spotify-style squads, OKRs and scaled agile.
  • Brought the new product framework to 90%+ adoption across the portfolio, with employee NPS up 7 points year over year.
  • Delivered 500+ hours of hands-on product training to 1,000+ employees, including managers from across the portfolio, in two- to three-day certification programs.

Fiserv

Product Manager 2018 to 2019 · Warsaw

Payments and transfers in a PSD2-regulated environment.

  • Owned product vision, roadmap and backlog, turning compliance requirements (Strong Customer Authentication, KYC/AML onboarding) into shipped features.
  • Led delivery of developer-facing Open Banking APIs for third-party providers under PSD2, and the web portals around them, with global engineering, QA and compliance teams.
  • Drove an agile transformation across departments that raised sprint commitment attainment (say/do ratio) by 30%.

10Clouds · Trust Stamp

Product Manager 2017 to 2018 · Warsaw

Led product for Trust Stamp, an AI identity-verification platform: facial biometrics and large-scale data mining to prevent fraud and support KYC/AML across fintech, law enforcement and real estate.

  • Managed full-cycle delivery of custom software and crypto solutions for three international startup clients.
  • Owned $500K to $1M budgets and led a 20-person delivery team across Europe and the US.

Orange

Project Manager 2014 to 2017 · Warsaw

Delivery of complex, cross-team initiatives at Orange Polska, reporting to program management.

  • Built and kept the project plans that gave program leadership accurate progress tracking and forecasting.
  • Managed dependencies and removed impediments across teams, and delivered on time and within budget.
  • Ran the Scrum ceremonies and the team's Jira setup, which made the backlog and reporting visible in real time.
3running right now
24/7live trading
10years product leadership
AustinTexas, US
How it looks up close

When the vendor is the bug.

A client's dashboard showed one contract open. His broker's own platform showed eleven. Somebody was wrong, and the only way to find out who was to stop guessing.

The obvious suspect was our code, so that is where I started — and four of the five contracts reconciled to the lot. That is the point where most investigations declare a rounding problem and move on.

Instead I put a line in the log next to the number, recording what the position feed actually returned each time it was read. Fifteen seconds apart, for one unchanged position, the vendor's API returned two different quantities. One row. Nothing skipped.

That turned an argument into a report the broker could not wave away, and it changed what we built next: the platform now computes the net position from our own recorded fills, and treats the vendor's number as a claim to be checked rather than a fact to be trusted.

A number you cannot reproduce is not a measurement. It is a rumour with a timestamp.

production log — redactedpositions feed
12:03:25  positions  acct ••••  •••  net=18.0
12:03:40  positions  acct ••••  •••  net=1.0

          rows emitted ....... 1
          rows skipped ....... 0
          our fills tape ..... unchanged
Same position, fifteen seconds apart, from the vendor's own endpoint.
Process

You approve the plan, not the pull request.

The point of every step below is that you can tell, without reading code, whether the thing is on track.

01

Frame

One page before anything else: what changes, for whom, and how we will both know it worked. If we cannot write that page, the project is not ready and no amount of building will make it ready.

02

Plan, then code

A written plan goes first — the files, the order, the migrations, what could break. You approve that. It is the last moment where changing your mind is free, so it is the moment worth spending.

03

Build in the open

Small branches, screenshots, and a running preview you can click today. No month-long silences and no demo that only works on my machine.

04

Prove it

Every fix begins with a test confirmed failing, so a passing test means something. Claims quote the log rather than paraphrase it. Where something cannot be proven, that gets said in those words.

05

Operate

Deploys, monitoring, migrations, the on-call. Anything that moves money or touches a customer stays behind a human gate — automation gets to prepare the action, a person still takes it.

Questions people actually ask

The awkward ones, answered first.

If the answer below is a problem for you, better to find out now than in week six.

Isn't AI-written code a risk?+

It is, if nobody reads it. So the gates are the product: a written plan before code, a test confirmed failing before a fix, an adversarial review whose job is to break the change, and a human reading every diff before it moves. Speed comes from running those steps in parallel, not from skipping them.

Who owns the code and the accounts?+

You do, from the first commit. Your repository, your cloud, your keys, your vendor contracts. If we stop working together the system keeps running and another developer can pick it up — that is a design requirement, not a courtesy.

What does it cost?+

Fixed price per feature when the scope is genuinely clear, hourly when it is not. I would rather quote a fixed price on a small first slice than a large number on a vague one, so the first engagement is usually deliberately small.

Do you trade money for clients?+

No. I build and operate the platform; every position-level decision belongs to the account holder. Exchange access on my side is read-only, and I do not give investment advice — the trading work here is software engineering, not money management.

What don't you do?+

Strategy decks with no build attached. Staff augmentation where somebody else owns the outcome. And any project where nobody can say what "done" means — I will help you define it, but I will not start building before we have.

Where are you, and how do we work together?+

Austin, Texas — US Central time, authorized to work in the US, and a decade of it spent in European delivery teams, so a spread-out timezone is the normal case rather than the exception. Written updates by default; calls when a decision needs one.

Bring the project
that keeps slipping.

Vague requirements, messy data, everyone agrees it should be automated and nobody has started. That is the one worth doing.

Straight to bart@exegov.ai. No list, no newsletter.