2026-07-11
The new definition of pre-seed
Mohamed BadwiiHead of Investment, Falak Startups
Badwii revealing that pre-seed requires inital traction, revenue and a clear path to market from day one..
Engineering studio · since 2017
We fix or build your product in a week.
No vibe coding. No AI slop. One bounded scope, one fixed price, one week. It starts with a paid audit — and sometimes the audit says don't build it.
19 industriesSince 2017$32M+ raised by our clientsToptal Top 3%
Clients
Backed by Y Combinator, ElevenLabs, CapitalG, Kleiner Perkins, Union Square Ventures, Anova Investments, Silicon Badia, and the Asian Development Bank.
Start here
Which of these is you?
Most people who call us are mid-failure, not mid-plan. Something broke, something stalled, or a date is coming.
What we do
Four ways to work with us.
Fixed scope, fixed price, fixed timeline. Nothing here runs longer than a week. If the work needs more than one week, each week is scoped and committed separately.
Product Audit
2–5 daysA written verdict on what is actually wrong, what it costs you today, and what to do first. Sometimes the verdict is: do not build this.
from $2,000Product Rescue
1 weekA build that someone else left broken, working again. Against a defect list agreed before the week starts.
from $6,000The Blocker
2–5 daysThe one technically hard thing that has stalled your team for weeks, solved and shipped in your codebase.
from $3,000Product Launch
1 weekOne real, working thing you can put in front of a customer, a community, or an investor. Built in a week.
from $6,000Why a week is possible
We are not fast because we rush. We are fast because most of it is already built.
Data access, background work, concurrency, pipelines, and auth run on patterns we have solved and reused for a decade. A build carries a fraction of the code it would elsewhere.
01
Scope is cut before the clock starts
A week only works if what enters it is honest. We scope down until the work genuinely fits, and we name out loud what is being deferred instead of quietly dropping it later.
Written before we start
The defect list, or the definition of done, is agreed in writing before the week begins.
Nothing enters mid-week
New scope becomes next week's decision, priced separately, or it does not happen.
Finished or not finished
No partial credit. At the end of the week it either works or it does not.
02
Built to change afterwards
Less code means fewer places a change can reach, and fewer places a bug can hide. That is what makes the second week cheap, and the pivot cheaper still.
Solved once, reused
Your budget goes into your product rather than the plumbing underneath it.
Swappable
Models, integrations, and whole flows slot in and out without touching the platform around them.
Yours to keep
You get the code, the reasoning, and enough of a picture to hand it to anyone else.
Where the cap came from
We did not always work this way.
We went back through eight years of engagements and sorted them by how they ended. Short, bounded work succeeded almost without exception. Long, open-ended builds failed, and none of those failures were technical. They were delivery failures.
A management system that took a year. A landing page abandoned at ninety percent. A rebuild that produced nothing in three months and cost us the client. The honest conclusion was structural, not motivational: the open-ended commitment is the failure mode.
The cap is the fix, and publishing it means it is measured in public. That is the point.
Capabilities
What we build.
The technical ground we cover. Any of it can be the subject of an audit, a rescue, a blocker, or a week.
AI agents & retrieval
Agents that read your data and answer from it, grounded in a knowledge graph rather than a guess. Swapping the model later is configuration, not a rebuild.
—Agentic RAG over your own documents and data
—Knowledge graphs and semantic search
—Custom NLP and entity extraction
—Tool use and multi-step orchestration
Custom neural networks
Production models built for your data and your problem. Not a prompt wrapper around someone else's API.
—Models trained on your own data
—Forecasting and graph neural networks
—Lipsync, audio, and vision networks
—Evaluation and monitoring in production
Workflow automation
The manual handoffs eating your team's day, running end to end. When the process changes, the automation changes with it.
—Sales, ops, and approval workflows
—WhatsApp and messaging automation
—Custom pipelines and orchestration
—Steps reordered or replaced without a rewrite
Data & knowledge graphs
The data your competitors pay a fortune for: scraped, structured, and kept current.
—Distributed scraping across hundreds of sources
—Entity and relationship extraction
—A live, queryable knowledge graph
—New sources added without reworking the pipeline
Product engineering
Full-stack delivery on the product itself: web, mobile, payments, and the architecture underneath.
—Web and mobile applications
—E-commerce and payment infrastructure
—On-chain systems and wallets
—Migrations off platforms you have outgrown
Infrastructure & DevOps
The deployment, monitoring, and infrastructure that keep a build fast and safe as it grows.
—Infrastructure as Code and CI/CD
—Environment and deployment automation
—Monitoring, logging, and alerting
—Security and data isolation
The work
Case studies
What we built, how long it took, and what happened after.
Cybersecurity2021
2 days
BreachSecureNow
Diagnosed a production failure across 32 microservices in 2 days, then stayed on for years.
First working result2 days
Stayed with us2+ years
Read case study->
Defense2025
2 weeks
Saudi Armed Forces, Quality Corps
Delivered in 2 weeks what was scoped for 2 years. The Quality Corps of the Saudi Armed Forces, off paper.
First working result2 weeks
Stayed with us2 weeks
Read case study->
Automotive AI2024, 2025
80%
Cybertron Labs
Automated the V-model for the first safety-aware engineering AI serving German Tier-1 OEMs.
First working result72 hours
Stayed with us16 months
Read case study->
AI & Media2022
YC-backed
Wondercraft
Built the MVP and the neural networks behind a YC- and ElevenLabs-backed AI audio studio.
Read case study->
GovTech2023
10× cheaper
Civi Digital
Automated what competitors sold at $20K per month on two-year contracts. Disrupted the market.
Read case study->
HR-Tech2020, 2021
Pre-ChatGPT
Tarjama / Ureed
Built the NLP-powered candidate matching behind the Middle East's largest freelancer marketplace, in 2020.
Read case study->
Interviews
Insights
Conversations with founders and investors building in Egypt.
"Enigma has been a true partner, solving our toughest challenges and accelerating our progress. When we needed custom AI, they delivered flawless results in days. We trust Enigma for high-stakes innovation.
Ahmed AbdienCEO and Founder, Cybertron Labs
Who you work with
Team
Muhamed Fahmy
Founder & Lead Engineer
10 years in automation. Toptal Top 3%. Ex-Tarjama lead.
Eike B. Post, PhD
Strategic Advisor
Berkeley, Stanford. 4 companies across 4 countries.
Ben Sheppard
Growth Advisor
4x founder, 1 exit. Speaker at Google HQ, the UN, the WEF.
Aline Avellar
Head of Projects
Università di Bologna. International delivery across 4 countries.
Start with the audit.
Two to five days. You get a written verdict on what is actually wrong, what it costs you today, and fixed scopes for the work. If we are not the right team, the audit will say that too.
