This is the closest match Fauzul has to anything at Lovable, because Forward Deployed Engineering is close to the job he has already done for nine years. He spent them inside other people's businesses: a Norwegian fintech pod he staffed and ran until Vipps payments were live in production, an engineering seat next to Skytracks Canada's technical CTO, in Zürich taken from a blank canvas to its first market release, and 's HIPAA clinical platform. The pattern was the same every time: take a fast-moving customer's real problem, ship something that works inside their business rather than in a demo, then keep changing it once real usage arrived. He builds full-stack in React, TypeScript and Go, and he has already put AI inside a customer's daily workflow: , a 26-year-old export house with no developers of its own, now changes its live site by saying what it wants in natural language. London is the seat he is ready for now, with a sister already there and a self-sponsored UK Global Talent route in progress.
Forward Deployed Engineer•London · Boston · New York · San Francisco•View posting ↗
Why he's a real fit
Three things worth reading in full.
01
Nine years of client-embedded delivery, which is what this role actually is
Hired and led a dedicated engineering pod from zero to production for Norway's Frontgo, shipping a Vipps mobile payments integration into their core financing systems. Built as the engineering counterpart to Skytracks Canada's technical CTO. Took in Zürich from blank canvas to first market release across a relationship that ran more than six years, until Loopcloud acquired it in 2026. The title Forward Deployed Engineer is new to him. The job underneath it is not.
02
He has already deployed intelligence into a customer's daily workflow
is a 26-year-old export house that had to call a developer for every brand change. Instead of handing them a CMS to learn, he made the change request itself natural language: someone on their team says what they want, an agent writes the pull request, the pipeline puts it live, and every edit stays reviewable and revertible in Git. That is the JD's own one-line summary, deploy intelligence into the real world, already done once for a customer who does not build software.
03
Agentic AI as daily practice, not a stretch for the posting
is an agentic due-diligence platform he architected and shipped solo, live at wiregent.com. Agentic payments connect to a double-entry, hash-chained ledger and a token economy, and its temporal certification is built to show a company changing and growing over a period rather than looking correct on one day. The diligence agents work a company's records the way a professional works a file: the verdict has to point at the evidence behind it and say out loud what it could not score, which is the only way an automated answer earns trust with a customer. First working platform from scaffold in eight days, agent-led throughout. applies the same pattern to visa eligibility, six country agents evaluating one profile in parallel.
The FDE mindset, in eight proofs
Hands-on, customer-embedded, agentic by default. A pattern rather than a résumé line.
Every line below has a shipped artifact behind it.
MF·01
Embeds with the customer
He has spent nine years inside other people's businesses, not advising from outside them.
Ten client geographies carried at once, remote-first since 2020, from Dhaka to Zürich to Oslo to Quebec. On he built browser audio editing with WebSocket collaboration inside the client's own stack, alongside their engineering leadership. The work happened in their repository, their release process and their constraints, which is a different job from writing a recommendation and leaving.
MF·02
Ships the agentic system, not just the pitch
Agentic AI is his daily practice, not a stretch for this posting.
Repeating workflows get codified into reusable agent skills; hooks and tools go in where they cut cost and tokens. The harness is chosen per task rather than per habit, which is why the chat on this page runs on-device instead of calling out to an API for a question about a résumé.
MF·03
Rapid prototyping under real constraint
He learned to ship in days before he could afford weeks.
He keeps entering timed builds in unfamiliar stacks: Bolt's World's Largest Hackathon in May 2025, the AI Agent Olympics at Milan AI Week, the Google × Kaggle agents intensive, and an EthGlobal Road to Web3 build in Solidity. Each one ended with something deployed and public rather than a slide about what it would have been.
MF·04
Product, code and architecture, one person
The same person scopes it, designs it, and ships it.
Designed 's brand, prototyped the menu with generative AI, and shipped the live ordering platform himself on a Go backend. Architected and built solo, from ledger design to deployment. This page was designed and coded by the person it argues for.
MF·05
Zero-to-one across five verticals
New markets and new problem shapes do not slow him down.
Five industries, and in each one the first version of the product was the one he shipped. Music-tech clients in Zürich, a fintech pod in Norway, clinical software in Dhaka, a restaurant in a new market, an esports league with no precedent locally. Same engineering standard held in every one.
MF·06
Systems that have to be right
The platforms underneath his client work were the ones that could not quietly fail.
's tenancy boundary, which every product inherited, on Auth0 with multi-tenant RBAC. 's HIPAA clinical pipeline, where a wrong record is a patient problem rather than a bug. 's double-entry, hash-chained ledger, where an entry that cannot be traced is worthless. Compliance and correctness designed in at the start, not retrofitted after an audit.
MF·07
Founding and early-startup environment
This is not a new environment for him. It has been the only one.
Co-founded in 2017 and grew it to 22 engineers with nobody handing him a quarter's scope. and are both zero-manager builds today.
MF·08
Multilingual, literally
English and Bengali fluent, Hindi verbal, and French in public every day.
A 900+-day French streak kept in the open, with grammar explainers published on YouTube and Instagram. Consistency as a habit, not a highlight reel.
What Lovable is looking for
Nine asks, mapped one for one. Tap "the bridge" on any ramping card.
The JD puts the job in one sentence: deploy intelligence into the real world. A founding FDE seat at the intersection of engineering, product, customer success and research, building with Lovable's most ambitious customers and feeding what you learn back into the core product. Tap any Ask Fauzul chip to go deeper than the summary.
REQ·01Strong
5+ years of experience building and shipping full-stack products
Nine years co-founding and running , hands-on first as UX Engineer and then Principal Engineer, later Engineering Manager and CIO, shipping 0→1 platforms across five verticals for clients in ten geographies. Since stepping out of the operating role in August 2026 he has kept shipping solo: 's ordering platform runs on a Go and ConnectRPC service he wrote and deployed himself, and is a full agent platform he architected, built and now operates alone.
REQ·02Strong
Proficiency in modern frontend (React/TypeScript) and robust backend development
React and TypeScript on one end, Go and Node on the other. is his solo full stack, a React and Zustand local-first frontend over a Go and ConnectRPC service on Cloud Run. 's production pipeline is Go. is a TanStack Start frontend over Python gRPC services behind a TypeScript GraphQL gateway. Behind those sit nine years of client work, from 's NestJS marketplace with multi-tenant RBAC on Auth0 to , where he wrote the first React and NestJS product and admin panel himself before standing up the team around it.
REQ·03Ramping
Strong architectural thinking and familiarity with distributed systems
He architected multi-tenant SaaS with RBAC and SSO at , ran event-driven integrations across client platforms on AWS and GCP, and designed 's double-entry, hash-chained ledger sitting under agentic payments and a token economy. The architectural thinking is his own work, not something he supervised.
The bridge
Those systems were built for tenancy, auditability and correctness rather than for millions of concurrent consumer sessions. The design instincts transfer; the traffic profile is the part he would be learning on the job.
REQ·04Strong
Ability to rapidly prototype, experiment, and learn from real usage
was built solo overnight during the World's Largest Hackathon in May 2025 and submitted to Y Combinator; it was not accepted, and he kept building. Enterprise operators trialling it are the reason the AI orchestration is now moving off the client and into the backend, so private models and enterprise security can sit behind it. That is real usage rewriting an architecture. went from idea to a production agentic pipeline in a one-week sprint for Milan AI Week, and , a six-agent visa eligibility mesh, was a solo Google × Kaggle capstone.
REQ·05Strong
Exceptional product intuition and ability to work directly with users
Nine years of direct client work across ten geographies, most of it spent with the person who owns the problem rather than a ticket describing it. At he built the admin workflows the operations team needed to keep a growing catalogue honest, not just the customer-facing product, and by 2023 his seat there had become the client and business relationship itself. 's HIPAA clinical platform, which he architected, now carries 64 doctors, 3,242 patients and 13,340 schedules across 7 centers, and he aligned it with Bangladesh Digital Health standards through a Ministry-level rollout. Reading what a customer needs before they can specify it is the skill all of that rewarded.
REQ·06Strong
Comfort operating in high-ambiguity, high-agency environments
Nine years co-founding and running with nobody handing him a scope, growing it to 22 engineers serving ten client geographies at once. He stepped out of the operating seat in August 2026 and is fully self-directed again: and are both zero-manager builds where the only structure is the one he sets.
REQ·07Strong
Bonus: Worked on agentic AI systems or built with LLMs before
Daily practice rather than a stretch to meet a bullet point. is an agentic diligence platform he shipped solo and runs live at wiregent.com, agent-led throughout the build. parses a candidate profile once, then evaluates it in parallel against six countries as independent FastAPI agents behind a Google ADK orchestrator. He codifies repeating workflows into reusable agent skills and adds hooks and tools where they cut cost and time, and he picks the harness per task: Claude leads client delivery, scripts with Kimi via OpenRouter and voices with Gemini TTS, and this page's own chat runs on-device.
REQ·08Strong
Bonus: Led zero-to-one product development
Shipped 0→1 across five verticals: music tech, gaming and esports, hospitality, health tech, and professional networks. began as a blank canvas he turned into a first web product, then a team, then a first market release, and Loopcloud acquired it in 2026. went from four validation pop-ups to a live ordering platform. went from a scene split across Discord servers to the first nationwide league run through the platform.
REQ·09Strong
Bonus: Worked in a founding or early startup environment
Literally ex-Founder. started in 2017 with him as one of its co-founders, structuring the founding team and its engineering values before it had customers, and he carried it through Principal Staff Software Engineer, Engineering Manager and CIO to 22 engineers. He handed over the operating seat in August 2026 and remains a stakeholder and advisor. , and each started from the same empty repository he started from.
The signature moment
He'd build this in Lovable, day one.
Ask him to build the workspace for Forward Deployed Engineer. Here is what ships first, generated the way Lovable generates everything else.
lovable.dev/build / forward-deployed-engineerLive
Build the day-one workspace for the Forward Deployed Engineer seat.
Building your workspace…
Cadence board
KPI dashboard
Hiring tracker
Monday leadership syncWeekly
Ran this shape for nine years at ELO, across ten client geographies and five time zones.
Wednesday priority re-checkWeekly
The JD's own example: Monday's priority reframed by Wednesday. This is the meeting that catches it before it costs a sprint.
Leadership offsite prepQuarterly
Agenda, pre-reads, and the follow-through system that makes decisions survive contact with the week after.
All-hands draft reviewMonthly
Shaping how the CEO communicates, which is the half of this role nobody sees.
Engineering: Cycle time
Nine years co-founding and running ELO, hands-on first as UX Engineer and then Principal Engineer, later Engineering…
People: Headcount plan
React and TypeScript on one end, Go and Node on the other. Haiba is his solo full stack, a React and Zustand…
Delivery: Compliance
TagRamp was built solo overnight during the World's Largest Hackathon in May 2025 and submitted to Y Combinator; it…
Executive scorecardDesign
Set the hiring bar at ELO and held it while scaling from zero to 22 engineers.
Panel → onsite → offerRun
Ran the loop end to end, including the hard conversations at the offer and headcount stages.
Post-offer landingOwn
The part most trackers skip: whether the leader actually lands, and stays.
Not just the mockup, the habit
How I actually run agents
Claude-led for client delivery. Everything else is deliberate experiment: skills, hooks, harnesses, and loops that cut cost, tokens, and time-to-ship.
Skills from repeating patterns
When a workflow shows up more than once, I turn it into an agent skill with reusable instructions and triggers, rather than a one-off prompt.
Hooks, tools, and plugins for cost and speed
Most builds add hooks, tools, or plugins aimed at lower spend, tighter token use, and faster delivery. It is deterministic control in the places where prompts alone waste budget.
Claude ↔ Antigravity CLI for images
Recently connected Claude to Antigravity CLI (agy) so image generation rides an existing subscription, iterating taste-of-design and sharpening a design skill through that loop.
Harness and model fit per task
Claude Code for primary client build work; Antigravity and others when the task or token budget asks for it. This site's chat runs on-device Gemma 4 E2B (WebGPU + MediaPipe, model hosted on R2). NewScriber scripts with Kimi via OpenRouter and voices with Gemini TTS via Azure.
The honest ledger
Three real gaps, each with the actual bridge.
The three US offices on this posting would need sponsorship; London does not
London ready now, no visa dependency on Lovable
Worth naming which of the four locations this is for. London is the one, and it is a plan rather than a preference: he is actively pursuing UK Global Talent, which is self-sponsored and progresses independently of any single hire, so a London seat puts no visa dependency on Lovable. His sister lives there. Boston, New York and San Francisco would need US sponsorship he does not hold today. He travels to customers wherever they are; the only question here is where the desk sits.
Production Rust experience
Short ramp from Go
Lovable's backend pairs Go with Rust. The Go half is shipped and running, in 's ordering service and 's pipeline; the Rust half he would be starting. The pattern he trusts for that is the one that has worked before: learn the language by shipping something in it, the way Solidity, Flutter and Python each entered his stack through a real build rather than a course.
Distributed systems at consumer-hyperscale volume
Different vantage, real systems muscle
His distributed-systems depth comes from multi-tenant, compliance-heavy platforms: 's multi-tenant RBAC marketplace on Auth0, 's HIPAA clinical pipeline, and 's double-entry, hash-chained ledger carrying agentic payments and a token economy. That is a different load profile from millions of concurrent consumer users, and it is the part of this stack he would be learning rather than teaching.
Beyond the role
The part the checklist cannot score.
Every requirement above is a box a strong engineer could tick. These are the reason to hire this one, and none of them appear in the posting.
01
He has already been an FDE, just without the title
Frontgo, Skytracks, and were the same shape of work this role names: sit with a customer, work out what they actually need, and ship into their business rather than around it. The title is new. The muscle is nine years old.
02
He builds the tools he proposes
Most engineers at this seniority spec agentic tooling and hand it to someone else to build. He ships it himself, solo, in days: 's audit runner, ' six-agent mesh, 's plain-English CMS, this page's own embedded chat. There is no translation layer and no waiting on capacity.
03
He leaves the customer able to run it
The measure of a deployment is what happens after the engineer leaves. 's team edits their own catalogue without a developer. Five client organisations learned GDPR, HIPAA, VAT and WCAG 2 alongside the software, so compliance stayed theirs. Shipping fast into a regulated customer's stack is only worth it if it does not leave them a debt.
04
Rituals and playbooks are already his instinct
Nine years of running a remote-first engineering org means he has built onboarding, delivery cadence and hiring processes from scratch more than once, and coached engineers into roles beyond the ones they were hired for. Designing how Forward Deployed Engineering works at Lovable, its rituals and org design, is a problem shape he has already solved for other functions.
05
Consistency you can audit
A 900+-day French streak, still running, with the grammar explainers published under his own name rather than kept private. An FDE seat is largely made of unglamorous returns to the same customer problem, and that habit is visible in public over years rather than asserted in an interview.
Frequently asked, honestly answered
The questions an interviewer would actually ask.
The candid answers, including the two things he would be picking up here rather than bringing: production Rust, and consumer traffic at hyperscale. Here is what is real instead.
You have never held the Forward Deployed Engineer title before. Why trust you to help build the function instead of just filling the seat?
Because the shape of the work is not new to me, only the title is, and at Lovable the title is new to everyone. Nine years at meant sitting inside a customer's build, working out what they actually needed, and staying until it ran in their business. Frontgo, Skytracks, , and were all that job under a different name. I would rather help design the rituals and playbooks for this function than inherit someone else's, and I have done that ground-up work before: 's own engineering practices, onboarding and delivery cadence started as a blank page in front of me.
Rust is on your stack list and you have not shipped it in production. How fast can you actually ramp?
I have not shipped Rust in production. What I have is Go and TypeScript at real depth on both ends of the stack, and a track record of picking up new languages when a project needed them: Solidity for an EthGlobal build, Flutter for a client's mobile app, Python for data and agent pipelines. I learn a language by shipping in it, not by reading about it first, and I would want a real Lovable customer problem to do the same with Rust.
The role talks about distributed systems. Have you actually run something at the scale Lovable's customers will bring?
My distributed systems experience comes from multi-tenant, compliance-heavy platforms rather than consumer apps serving millions of concurrent users. 's multi-tenant RBAC marketplace, 's HIPAA clinical pipeline, and 's hash-chained ledger with an always-on event stream are the systems I have actually run in production. The design work transfers, the traffic profile does not, and that is the part I would be learning from the people here who have already done it.
You have run a company. Why would one of the first FDE seats at Lovable, embedded in someone else's roadmap, appeal to you more than building your own thing?
I am already building my own thing, , and I like that it is mine. But nine years of my career were spent doing exactly what this role asks: sitting with a customer, absorbing their ambiguity, and shipping something that worked inside their business. That was not a lesser part of founder work for me. It was the part I was best at and enjoyed most, without the parts of running a company that had nothing to do with building.
Why London and not Boston, New York or San Francisco?
London is where my ties are real. My sister lives there, and it is the city on this list I am ready to move to now rather than in the abstract. It is also the cleanest one operationally: I am pursuing UK Global Talent, which is self-sponsored, so hiring me in London does not create a visa dependency, while the three US offices would. I have not worked in Boston, New York or San Francisco, and I will not pretend otherwise.
This is a brand-new function with no playbook yet. How do we know you will not need a lot of structure to be productive?
Look at what I am building right now with nobody assigning it. went from a scaffold to a working platform in eight days, alone and agent-led, and I have kept shipping against it since; it is live at wiregent.com. Nine years running with nobody scoping my week is the same shape at higher stakes: ten client geographies running in parallel, and the only cadence in the room was the one we built. Ambiguity has been the environment the whole time, not an exception I need coaching through.
9Years co-founding ELO
22Engineers led
10Client geographies
Frontgo pod: zero to Vipps-integrated productionGlobal Jute: plain English in, shipped pull request outEKAGRA: 64 doctors · 3,242 patients · 7 centers10 client geographies, 5 verticals, 0→1 every time
Ready to build Forward Deployed Engineer?
Genuinely ready to work from London full time, on the CEO's clock.
REF · LOV-FDE · Forward Deployed Engineer · London
