Nine years of software. One clear conviction.
codestra was founded in Cologne in 2017 by Sajad Ghawami. Nine years of work on platforms for insurers, banks and industry has grown into one clear specialisation: AI systems for mechanical and plant engineering that hold up in daily operation.
Three convictions that shape the work.
Success is measurable - or it was not success.
We work against agreed metrics, not against impressions. If the effect cannot be quantified, we leave the project alone.
Respect for what is already there.
Legacy systems exist because they worked. Our job is not to throw them away - it is to decide which parts stay, which go, and which come later.
No dependency.
Our business model is not selling you a proprietary black box. It is getting the next project because the first one worked - and because your people can run it themselves.
Who is behind it.
Sajad Ghawami
Architecture and development at the intersection of fullstack, cloud and data platforms - for insurers, banks, industry and the public sector. Focused since 2023 on production AI systems in regulated environments, today on mechanical and plant engineering: making technical documentation, service reports and drawing archives answerable.
Senior engineering from first day to handover.
Projects are owned by experienced engineers, not by a layer of junior consultants. No subcontracting, no black-box delivery. What you are meant to run, we hand over running - documentation and operational handover included.
Nine years, briefly told.
Founded in Cologne.
Started with software and web development for clients whose external consultants had delivered more slides than running code.
Platform frontends for regulated sectors.
Rating and sales journeys for insurers, self-service tools for banking. The first larger team setup, and the first production systems in daily operation.
From the frontend down to the data foundation.
Pure UI mandates grow into complete platform projects on the client side: data models, backend APIs, cloud infrastructure. The profile sharpens away from advisory work and towards deliverable code running in the client’s operation.
The turn towards AI platforms.
The first production AI hub platform for a heavily regulated industry. RAG pipelines, vector databases and LLM observability become part of our standard stack.
AI platforms for regulated industries.
Design and implementation of client-specific enterprise AI platforms in banking, healthcare and industry - on-premise capable, multi-model and multi-cloud, with native Microsoft 365 integration.
Technical documentation as the data foundation.
Hybrid RAG, GraphRAG and natural-language querying become part of the standard delivery - not for their own sake, but to make operating manuals, service reports and bills of materials answerable. The projects with the clearest benefit come from industry.
Specialising in mechanical and plant engineering.
The industrial projects become the focus: service, sales, design and technical documentation in companies whose knowledge has grown over decades. The same engineering discipline as in banking and insurance, applied to machinery instead of policies.
Are we a fit?
Probably - if you put working AI ahead of good-looking slides.