I'm a Technical Project Manager with 8.5 years as a backend developer, now managing delivery teams at BetaCraft. I build the AI tooling my teams run on, and I read the codebase before I scope the work. Open to senior technical delivery roles in Europe, remote or relocation.
// Coder // Rider // Photographer // CookTwo internal systems I built and maintain at BetaCraft.
One covers the development workflow from brief to handoff. The other automates PM operations from call transcript to Jira ticket. Both are built with Claude Code and MCP, and both run on live client engagements, currently four.
Engineering workflows for the full dev lifecycle, from first brief to client handoff.
Automates the PM admin layer, from client call to Jira ticket.
Both systems are internal to BetaCraft. I built them and maintain them.
I wrote code for eight and a half years before switching sides.
I was a developer first. That changes how I manage.
I started as a developer because I wanted to know how things work. After eight and a half years of backends, APIs, and system architecture, you stop seeing software as code and start seeing it as a series of decisions, trade-offs, and consequences that someone has to own.
What pulled me toward project management wasn't the frameworks. It was noticing that most of the risk on a software project sits in the gap between what the client asked for and what the engineer built. I'd spent years on the engineering side of that gap and wanted to work both.
Now at BetaCraft I build the internal AI tooling our delivery runs on, alongside managing four client engagements that span enterprise integration work and AI platforms.
No PM-shaped proxy between you and engineering. I can challenge an estimate, spot an architecture risk, and still write the client update, without a translator.
I built, documented, and shipped the vibe-* dev framework and the PM skills system myself. Both are in daily use.
Client delivery at BetaCraft runs from brief to sign-off under one person. Scope changes, blockers, and budget conversations come to me directly. There is no escalation layer above me, so when something slips, it is mine to own and mine to fix.
I spent 8.5 years in LAMP / Laravel, writing backends and connecting third-party APIs. I still read the codebase before I scope anything.
I build multi-agent workflows and RAG pipelines with Claude: systems that take actions, run checks, and feed outputs directly into delivery, with a human reviewing what goes out.
I run sprint planning and client communication, and I keep teams moving when the requirements aren't clear yet and the time zones don't line up.
I've built and delivered CRMs, ERPs, and platforms across e-commerce, energy, and manufacturing, designed to keep running after the launch week.
Six engagements across institutional research, LegalTech, fintech, and agentic AI. Each one is written up in detail: what the risk was, what I did, and what the evidence shows.
Fixed institutional deadline, a compliance gap mid-sprint, and a distribution channel blocked with 400 participants waiting, all in the same sprint. The date could not move.
Regulatory Compliance Platform · 2026Two technical teams with conflicting accounts of what had been built, and no shared ground truth. I ran the production API tests myself, surfaced six SOW discrepancies, and documented what the evidence showed.
Brieflex.ai · 2025–2026No sprint plan. No time tracking. No SOW for work already in flight. How I established what had been built, where the process had broken down, and got the delivery moving again toward commercial launch.
Stefanov Capital · 2026Marketing site, HubSpot lifecycle automation, and a third-party checkout handoff, all in four weeks. How the integration contracts between parties were defined and held.
BetaCraft Internal · 2025–2026A two-part AI-native delivery pipeline: 27 development workflow skills and 16 PM operations skills. A client email goes in one end; a tested, reviewed delivery comes out the other. The system that now runs BetaCraft's live client work.
Personal Product · BetaCraft · 2025–2026Built because the thing I needed did not exist. A macOS voice-to-prompt app, iterated to v2.4.1 across eight specialist modes. No brief and no client, so every bad decision ships straight into my own daily workflow.
Six engagements at BetaCraft: four live now, two delivered. The live ones span an AI LegalTech platform, a South Korean social platform, a US healthcare compliance platform, and a three-system broadband integration.
US healthcare compliance platform. Three organisations could not agree on what had been built, with no shared record to settle it.
Case study →A marketing site, a HubSpot lifecycle automation layer, and a third-party checkout handoff, with the integration contracts between them defined and held.
Case study →Social platform for the South Korean market: social feed, stories, and event planning. Picked up mid-build after PM departure.
I picked up a struggling AI LegalTech platform mid-build, rebuilt the sprint process, established what had been delivered, and pushed toward commercial launch from a free-beta baseline.
Visit site ↗GPT-4o migration on a live chatbot platform: three parties (engineering, QA, client), six weeks, no service interruption.
Visit site ↗World Bank Group research platform: sole TPM, four-week fixed deadline, 400 enrolled research participants, and research-ethics compliance requirements with no flexibility on the date.
All projects managed at BetaCraft · 2024–Present
Selected projects from 8.5 years as a hands-on developer.
Multi-channel order automation: Amazon SP-API, WooCommerce, and Etsy, connected and running. Processing efficiency up 30%.
Integrated GPT content generation and GMB API automation into Pendio CRM. Campaigns using the generated content performed about 20% better; the GMB automation increased marketing output by roughly 25–30%.
ERP for an Australian diamond wholesaler. Core order and stock workflows ran about 50% faster after the rebuild; making inventory visible to customers lifted sales roughly 15%. I added two subsystems that weren't in the brief but that the workflows clearly needed.
CRM for solar and wind energy assets. Integrated Locus hardware APIs for real-time monitoring and billing.
CRM for a Copenhagen client, architected on AWS for a projected user base in the millions.
Ticketing and asset management system that made it easier to see what was broken and who was on it. Integrated Uptime Kuma for live status monitoring.
I spent ten years building things before managing them. The tools are the same.
I run projects Agile end-to-end: sprint planning, risk tracking, and stakeholder updates. The process serves the delivery, not the other way around.
I don't need a glossary between engineering and business. I need ten minutes with the ticket and the person who wrote it.
My engineering base is PHP and Laravel: backends, third-party integrations, and Swagger docs that cut developer onboarding time by about 25%.
I'm building agent workflows whose output goes directly into client deliverables, with a human review step where it matters.
I write about agentic AI delivery, EU AI governance, and what a decade of writing code looks like when you're managing the team that writes it.
Beyond the code.
I speak three languages (English, Hindi, Bengali) and switch between them depending on which city I'm in. When work's done, I'm usually on a motorbike somewhere, behind a camera, or making something questionable for dinner.
I visit new cities, often solo. Three languages give me three different ways of navigating somewhere unfamiliar, and getting lost is usually the best part.
I shoot street and travel photography, looking for the frame before the moment disappears.
I ride long highways with no fixed itinerary, preferably somewhere the signal drops before the fuel does.
I cook mostly improvised and rarely from a recipe, with strong opinions about what makes a proper meal.
Led end-to-end training for two consecutive PHP intern cohorts at WebCodeGenie, from onboarding through to live project contribution.
Built training modules and assessments from scratch, mapped to what the job requires rather than what a course outline says it should.
Taught interns how to communicate with clients: writing requirements, running calls, and handling the difficult conversations.
As Senior Engineer and Project Lead, I spent as much time teaching as I did building. Onboarding two intern cohorts end-to-end (designing the curriculum, running the sessions, getting them to a point where they were contributing to live projects) is where I learned that managing people is harder than managing code.
I'm looking for senior technical delivery roles, preferably the kind where reading the codebase is part of the job. Remote and international both work.