AI in practice
LLMs and ML with measurable value in real products — with evaluation and guardrails, not a demo that tips over in week two.
LLMs · RAG · Evals
// continuous delivery
AI in production — and everything underneath it.
Software Engineer · Freelance · Remote / Germany
RAG systems, agent-ready codebases and the infrastructure they stand on — designed, built and run by one senior engineer who answers for the whole thing.
Shipping since 2009
Full-stack · Infra · DevOps
Code → Cloud — one accountable engineer
Agent-ready codebases
// services
One engineer, the whole stack — wherever it gets hard.
Embedded as a senior in your team — or the person you bring in when it really matters.
Developer Experience
Tooling, conventions and clear default paths that make the whole team faster and happier — humans and AI agents alike. An effect that compounds with every working day.
Tooling · Automation · Standards
Full-stack development
Features end to end — from the database schema to the pixels users see. Typed, tested and built so the next person can maintain them.
TypeScript · React · Node · Go · SQL
Infrastructure & Cloud
Resilient, observable systems that scale predictably and tolerate failure. Infrastructure you can understand — and afford.
AWS · GCP · Terraform · Kubernetes
DevOps & Pipelines
CI/CD and release engineering that turn “works on my machine” into reliably shipped software every time — without late-night emergencies.
GitHub Actions · Docker · IaC
Technical consulting
Architecture reviews, second opinions and pragmatic roadmaps — from someone who has made these calls often enough to know how they play out.
Architecture · Audits · Strategy
// selected work
Outcomes, not output.
Two projects, documented in full — architecture, decisions, numbers.
CASE 01
sustaind · Compliance
Read the case study
CASE 02
sustaind · Developer Experience
Read the case study
sustaind · Compliance
Turning hostile documents into structured knowledge.
Architected and built sustaind's document-ingestion worker — a fault-tolerant pipeline that turns scanned PDFs, legacy Office files and huge spreadsheets into semantically grouped chunks, each traceable to the exact source pixels. Zero silent failures in production.
sustaind · Developer Experience
A codebase where AI agents engineer end to end.
Architected and drove the consolidation of seven repositories into one agent-ready Nx monorepo — three commands from clone to a running product on a local Kubernetes cluster, and one path to production instead of five.
// who you work with
I am Bastian Schöntag — a software engineer. I turn fuzzy problems into systems that run, and keep running.
Mostly I get called when a system has to land: a pipeline that keeps breaking, infrastructure that costs too much, an AI feature stuck at the demo stage. Those are exactly the problems I like: concrete, honest, solvable.
LocationGermany · Remote (EU)
ExperienceSince 2009
LanguagesGerman · English
AvailabilityOpen · Q3 2026
// let’s talk
Got a project that has to work?
Tell me what you are building and where it is stuck. I will tell you how I would approach it — a clear answer, no marketing, no obligation.