Five years in software, three of them full time on the backends behind Zameen.com, Bayut.sa and Propforce at 40M+ monthly users, then the AI layer above them. Now a Forward Deployed Engineer at UnternehmerTUM's Venture Lab in Munich.
Working in
Next six systems, each drawn as the architecture it actually is
Frankfurt am Main, Germany
MCP gateway
Zara AI
Venture Lab
Ask Ahmad
every backend
Verona research
rag-integrity-guard
Propforce CRM
benchmark API
retrieval
Zameen.com
Career OS
Career OS
CRM search
queues
production
deployments
CI/CD
Venture Lab
Social Mantra
Stacks contracts
thesis
live demos
Propforce
Monthly users served by backends I built and operated at DubizzleLabs
case study →Less PR prep time, from 40 minutes down to 6, via the agentic delivery pipeline
case study →Lower query latency from MySQL/PostgreSQL schema and index work
case study →Active clients on the end-to-end mortgage feature I shipped
case study →I spent three years at DubizzleLabs on the systems behind Zameen.com, Bayut.sa and Propforce: distributed Node.js/TypeScript services, transactional data at real volume, and the unglamorous work that keeps them fast (query plans, indexes, background workers, observability, RBAC).
Then I built the layer above. AVA, an internal AI copilot that takes a request in chat all the way to a reviewed pull request. MCP servers that give LLM agents typed access to internal APIs instead of engineers digging through docs. A RAG knowledge base over architecture decisions that answers in plain English. None of it was a demo. Colleagues used it every day.
I'm now doing an M.Sc. in Computer & System Engineering at TU Ilmenau, researching a data-integrity layer for LangChain QnA systems: blockchain-anchored fingerprints re-verified at inference time, so a poisoned corpus can't quietly answer for you.
Since August I've been a Forward Deployed Engineer at UnternehmerTUM's Venture Lab in Munich, which is the fastest version of this work I've done. I get dropped into a product team and the job is to get something usable out in days: an app for validating venture AI use cases, tooling for checking work against AI policy and EU AI Act criteria, internal apps that pull scattered org processes into one place.
What ties it together is simple enough. I want the thing I build to end up in someone's daily workflow, which is what I aim for at the Venture Lab and in the projects I build on my own time.
Five years of paid work, three of them full time. Study and research sit below, separately. The full career and the resume carry the detail.
2026
2024
2023
2022
2021
Education
M.Sc. Computer & System Engineering
TU Ilmenau · Mar 2026 – Oct 2028 expected
B.Sc. Computer Science
FAST NUCES · 2019 – 2023 · Dean's Honour Roll, ranked 7th nationally in the NU admission test
Alongside
Data-integrity research
University of Verona · 2024 – now, with two PhD researchers
Aspire Leaders Program
Incubated at Harvard University · 2023
Community Evangelist
Algorand UK · 2022 – 2024
Everything here appears in a system above. Nothing is listed because I read about it. Pick a node to trace what it was built with.
TypeScript every backend since 2023
AI and agents
Model Context Protocol · agentic pipelines · RAG · LangChain · embeddings and vector search · pgvector · LLM evaluation and guardrails · Claude, OpenAI, Gemini and Groq APIs
Backend
TypeScript · Node.js · NestJS · Express · Python · FastAPI · REST · GraphQL · microservices · background workers · RBAC and authentication
Data
PostgreSQL · MySQL · MongoDB · Redis · OpenSearch · Supabase · Sequelize · Drizzle · query optimisation · indexing and schema design
Infrastructure
AWS · Docker · GitHub Actions · CI/CD · Vercel · logging, metrics and tracing · Git
Frontend
React · Next.js App Router
Also written
Go · C++ · Solidity and Clarity smart contracts
Systems I designed, built and ran, each with the problem, the architecture, and what changed afterwards.
Fingerprints every RAG chunk at ingestion and re-verifies it at query time, so a poisoned or tampered chunk is caught before it reaches the model instead of the vector store being trusted by default.
Read the case studyCampaign orchestration for an autonomous voice agent: callbacks, follow-ups, task creation and WhatsApp follow-through across Propforce and Jarvis, with an on/off switch the business team operates without a deploy.
Read the case studyA chat request becomes clarified requirements, LLM-drafted wireframes, a filed Jira ticket, and a branch of generated code and tests opened as a pull request under human review.
Read the case studyInternal microservice APIs exposed to LLM agents as typed, permissioned tools, with dynamic schema generation, per-caller rate limiting and full invocation logging.
Read the case studyThe production backend behind Propforce, Zameen.com and Bayut.sa: transactional services, background workers, and an end-to-end mortgage flow adopted by 5,000+ clients.
Read the case studyBlockchain-anchored Merkle fingerprints over a retrieval corpus, re-verified at inference time, so tampering with source documents is detected before a poisoned answer is served.
Read the case studySix more, beyond the six above
Nine references, quoted in full and in their own words. The first names the work in the grid above it — pick a face to read any of the others.
Independent confirmation of the three Dubizzle systems above
“I had the opportunity to work closely with Muhammad Ahmad on several AI-focused initiatives, including building MCP-based integrations and agentic systems capable of calling, messaging, and engaging customers across different channels.
What stood out most was his ownership. Once Muhammad took responsibility for something, he could be relied upon to drive it forward, solve problems, and deliver without needing constant follow-up.”
Zeeshan Hasan
Product & Engineering Leader · managed me directly at DubizzleLabs
20 July 2026
Eight more, in their own words
I have been working with Muhammad Ahmad at Dubizzle Labs for the past 1.5 years, where we are part of the same team. Ahmad consistently demonstrates strong collaboration skills and a proactive approach to his work, contributing effectively to team success. He is reliable, detail oriented, and always willing to take on challenges to ensure high-quality results.
I have also known Ahmad since our time at FAST NUCES, where he was my junior during our Bachelor's studies. His organizational abilities were evident even then, particularly through his involvement in successfully managing several SOFTEC events.
I look forward to seeing Ahmad achieve even greater milestones in his career.
It is my pleasure to highly recommend Muhammad Ahmad as an exceptional individual and professional, particularly in the realms of technology and the Web3 ecosystem. I had the privilege of collaborating with Muhammad during the Stacks Hackathon, Stacks Clarity Developer Bootcamp and Stacks Xplore1.0 conference, where we delved deep into the intricacies of the Web3 landscape, and it was an enriching experience.
Muhammad is a tech enthusiast with a strong passion for creating impactful solutions within the technology ecosystem. His expertise as a backend engineer is evident in his meticulous approach to problem-solving and his dedication to delivering high-quality results. What sets Muhammad apart is his unwavering commitment to leveraging technology for positive change and his continuous exploration of Dapps, DAOs, and NFTs.
Beyond his technical prowess, Muhammad's leadership capabilities are truly commendable. His purpose-driven leadership style, coupled with his resilient mindset, makes him an invaluable asset to any organization. He approaches challenges with determination, always striving to find innovative solutions that drive meaningful impact.
Muhammad's mission to democratize decentralized solutions and his deep understanding of emerging technologies make him a standout professional in the tech industry. I have no doubt that he will continue to make significant contributions and inspire those around him.
I wholeheartedly endorse Muhammad Ahmad and am confident that he will excel in any role that allows him to harness his skills, passion, and vision for a decentralized future.
I had the pleasure of working with Ahmad in multiple capacities during our undergrad days. He is a passionate individual and a true go-getter. Once he sets his eyes on a target, he gets into a montage of dedication and focus to deliver the results.
The way he has enhanced his skill set on the technical side is truly remarkable, and it serves as the cherry on top for his excellence in social and leadership skills.
For anyone looking to work with Ahmad, I recommend him as an excellent software engineer and a formidable team player.
Muhammad was an intern at Educative Answers during the summer 2022. As an intern, he wrote 30+ articles and reviewed 40-60 articles written by other Answers interns. If I have to pick out an intern with an amazing work ethic, it'd be Muhammad Ahmad. He proved to be extremely hardworking, well spoken and detail oriented over the course of his Educative Answers internship. Muhammad Ahmad was assigned some of the really challenging topics in the Networking, Data Structures and the operating system domain. He manifested a remarkable attention to detail by creating answers that hardly required any edits. I think he'll be a wonderful addition to any workplace that he ends up in and has a bright future ahead.
Ahmad is a kind and intelligent person with whom I liked working. During his work, I was fortunate to be his Technical lead. I appreciate how precise and in-depth his work is. I wish him the best of luck in his future endeavors.
Ahmed is very talented person. He is a quick learner and always ready to participate in any work. His passion for learning new things is very admirable. Always a great pleasure working with him.
It's a pleasure working with him as he creates a positive environment around his positivity. I was lucky enough to review some of his work and I like how detailed and deep it is. Some of his answers were so good that I specially asked the community team to tweet them to the public.
Ahmad has great passion for learning things out of the box!
Written on LinkedIn, where each one can be read at its source. Two were sent to me directly.
All nine, laid out in fullThat is the part I like. Available for full-time roles, consulting engagements and team-lead work, in Germany or remote, now and after the M.Sc. Email is fastest, and I reply the same day.