Repetitive manual work
Reduce time spent collecting, classifying, validating, routing, and reporting information.
7+Years engineering experience
3Production products built
4Research publications
End to endFrom idea to deployment
Where useful AI opportunities are often hidden.
I focus on business problems where better workflows, automation, or AI can create measurable value.
Knowledge that is hard to use
Make information across documents and systems easier for employees or customers to access.
AI ideas without a clear path
Identify which opportunity is feasible, valuable, measurable, and worth implementing first.
Two focused paths from problem to production.
Choose the starting point that matches what your company already knows and is ready to implement.
AI Opportunity Snapshot
We examine one important workflow through two focused conversations and limited, non-sensitive examples.
You receiveA one-page opportunity brief within five business days after the second conversation, covering feasibility, required data, main risks, a success metric, and the practical next step.
A good fitOne recurring workflow, a process owner available for two conversations, and serious interest in implementation if the opportunity is validated.
AI and software product development
Turn a validated idea or prototype into reliable software that fits your users, product, and operating environment.
You receivePaid technical scoping, architecture, AI or LLM integration, backend APIs, data workflows, cloud deployment, monitoring, documentation, and handover.
A good fitA defined user and business problem, an involved founder or product owner, a realistic budget, and access to the necessary data or systems.
One delivery process, with the right starting point.
SMEs can begin with opportunity discovery. Startups with a validated idea or prototype can begin with paid technical scoping or implementation.
Fit call
Confirm the problem, users, owner, available evidence or data, and implementation intent.
Discover or scope
Use the free SME Snapshot or paid startup technical scoping to define the right solution.
Actionable plan
Receive feasibility, architecture, data needs, risks, success metrics, milestones, and the recommended next step.
Build if approved
Develop, integrate, test, deploy, document, and hand over the solution through an agreed paid engagement.
Two clear entry points: the free Snapshot is for qualified SMEs exploring one workflow. Startup product development is a paid engagement for validated ideas or prototypes.
Book a fit callSelected work with real delivery depth.
Concise proof here, with the problem, delivery decisions, and technical depth in each full case study.
MeydaNews
Live AI productFounded and launched a media intelligence product connecting collection, source analysis, grounded retrieval, APIs, and cloud operations.
Verified outcome: a working end-to-end product carries raw media feeds through to evidence-backed analysis and a review interface.
Quality-aware recommendation
Applied research systemDesigned and evaluated a recommender that balances personal relevance with independently validated signals of source reliability.
Verified outcome: average recommended-source quality increased by 3 points while baseline precision was maintained.
PhotoCans
SaaS productCo-founded and delivered a privacy-first product spanning customer journeys, administration, licensing, deployment, and releases.
Verified delivery: supported presentation, sales, onboarding, licensing, and support for more than 100 users across multiple production releases.
ReserveAfzar
Internal workflow systemDesigned and deployed a shared booking system for hall availability, recurring reservations, and administrative reporting.
Verified operational change: one shared workflow replaced informal coordination and made availability visible to staff.
AI expertise with delivery ownership.
Research discipline, production engineering, and clear ownership in one engagement.
01AI and software expertiseRAG, recommenders, FastAPI, AWS, evaluation, and MLOps.
02Proven deliveryThree products moved from requirements and architecture into production.
03Measured ownershipOutcome focused decisions, observable quality, documentation, and handover.
Organizations and products I have worked with
MeydaNews
University of Potsdam
Max Planck InstitutePhotoCans
Bu-Ali Sina University
“His critical thinking and programming abilities make him a skillful programmer who can quickly implement theory into practice.”
“He consistently impressed me with his ability to understand complex concepts and performed at a very high level.”
“He built a Reddit data pipeline from scratch and developed a novel recommender system, demonstrating exceptional technical skill.”
Start a conversation
Which workflow or product should work better?
Tell me about the operational problem, validated product idea, or prototype you want to improve. I will recommend the most useful next step.
✓ SMEs: one recurring workflow or operational bottleneck✓ Startups: a validated idea, prototype, or existing product✓ A decision-maker available with serious implementation intent