$ whoami

Martin Oputa

AI-Native Full-Stack Product Engineer · Berlin

AI-native full-stack product engineer with 5+ years building production systems in React, TypeScript and Node.js. I take products end to end, from first commit to paying users: I built and run PracticeDepth, a live AI interviewer on LangGraph, on my own. Background in real-time, event-driven systems (Kafka, WebSockets, WebRTC, MQTT) for EU emergency-response projects.

view work →get in touch
Martin Oputa
Martin Oputa
Berlin, Germany · GMT+1
available for work
martin@berlin — zsh
martin@berlin:~$ cat stack.json
{
  "frontend": ["React","Next.js","TypeScript"],
  "backend":  ["Node","FastAPI","Postgres"],
  "realtime": ["WebRTC","Kafka","WebSockets"],
  "ai":      ["LangGraph","RAG","MCP"]
}
martin@berlin:~$
0+
Years shipping production
0%
Database latency reduced
0%
Fewer production bugs
0
EU Horizon 2020 projects
·JavaScript·TypeScript·React·Next.js·Node.js·Python·FastAPI·PostgreSQL·Kafka·WebRTC·WebSockets·Redis·Docker·AWS·GCP·Vercel·CI/CD·Git·GraphQL·Prisma·LangChain·LangGraph·MCP·RAG·Vector DB·GenAI·LLMOps
capabilities

Agentic systems, automation & full-stack —
from database to production to monitoring.

01/05

AI Engineering

Production AI products: LangGraph agents, RAG, pgvector, MCP servers, voice (Whisper STT, TTS) and AI-native delivery with Claude Code.

02/05

Backend Development

Robust, scalable server-side systems & APIs with Node, Express, Nest.js, Python, FastAPI, GraphQL and PostgreSQL.

03/05

Frontend Development

Intuitive, responsive interfaces with clean, maintainable React, Next.js and TypeScript.

04/05

Cloud

Deploying and optimizing infrastructure on AWS, GCP and Vercel for reliability, speed and scale.

05/05

CI/CD

Automated build, test and deploy pipelines with GitHub Actions and Docker for fast, reliable delivery.

selected work

Featured projects

Production-grade AI, real-time & geospatial systems — click any project for more details.

personal projects
professional work
career

Work experience

Five years shipping production software, from live AI products to real-time, geospatial, and mission-critical systems.

AI-Native Full-Stack Product Engineer

PracticeDepth· Berlin, Germany

2026 — Present
Next.jsTypeScriptLangGraphGeminiPostgreSQLPrismaRedisStripeDocker
  • Designed, built and operate an adaptive AI technical interviewer on my own, taking it from first commit to production in under 12 weeks; processing live Stripe payments since August 2026.
  • Built the interview engine as a LangGraph.js state machine with Redis checkpointing that pauses and resumes across stateless requests; code, not the LLM, picks the next question, so the flow stays deterministic.
  • Added voice at the edges (Whisper STT, Gemini TTS) and shipped code challenges, CV Defense, and a support assistant grounded in RAG.
  • Building a B2B hiring-screen product: multi-tenant workspaces, role-based access, row-locked credit reservation, and an exam mode with integrity signals, behind a CI-enforced feature flag.
  • Own delivery and ops: Linear tickets dispatch parallel Claude Code agents on GitHub Actions that open CI-gated draft PRs (1,100+ tests, payment-code guard); self-hosted on Hetzner via Coolify, monitored with Sentry and Grafana.

Software Engineer · Full Stack & AI

Astrial GmbH· Berlin, Germany

2023 — 2026
ReactTypeScriptNode.jsPostgreSQLKafkaWebRTCWebSocketsGitLab
  • Developed real-time communication systems using React and WebRTC while collaborating with cross-functional teams, achieving low-latency video and chat for emergency response scenarios.
  • Implemented Kafka-backed event streams for real-time data flows, enhancing geolocation tracking, communication state updates, and event-driven architecture.
  • Took full ownership of an end-to-end drone operations platform from planning to deployment, using the DJI Cloud API and MQTT to integrate real-time telemetry with automated 3D mission planning and remote command execution.
  • Optimized database performance with Redis caching and Prisma ORM, reducing query latency by 30% for PostgreSQL services.
  • Contributed to architectural decisions, code reviews, and cross-functional planning, and delivered live technical and product demonstrations to EU stakeholders.
  • Worked as technology provider on EU Horizon 2020 projects including NIGHTINGALE, APPRAISE, SYNERGISE, and IFLOWS.

Software Engineer · Full Stack

Vub GmbH· Leipzig, Germany

2022 — 2023
Next.jsTypeScriptNode.jsPrismaJestCI/CDGitHub ActionsJira
  • Engineered internal tools and web apps using Next.js and TypeScript, integrating frontend clients with Java-based microservices.
  • Reduced production bugs by 25% within 6 months by implementing Jest testing suites with MSW and enforcing Clean Code / SOLID principles.
  • Mentored junior developers via code reviews and technical workshops on modern JavaScript architectures and Agile best practices.
  • Managed team priorities, deadlines, and deliverables using tools such as Jira to maintain strong Agile methodology and practice.

Software Engineer · Frontend

Posterlounge GmbH· Leipzig, Germany

2021 — 2022
JavaScriptjQueryReactMaterial UIUI / UX
  • Planned and executed the migration of legacy UI components (HTML with jQuery) to modern frontend frameworks (React), improving maintainability and consistency across the interface.
  • Created UI animations using CSS3 and JavaScript libraries such as Material UI and Bootstrap.
  • Translated designs and style guides provided by UI/UX into functional, pixel-accurate user interfaces.
toolbox

Skills & stack

JavaScriptTypeScriptReactNext.jsNode.jsPythonFastAPIPostgreSQLKafkaWebRTCWebSocketsRedisDockerAWSGCPVercelCI/CDGitGraphQLPrismaLangChainLangGraphMCPRAGVector DBGenAILLMOps
education

Education

2018 — 2020
M.Sc. Sustainable Business Development
University of Leipzig
2011 — 2015
B.Sc. Mathematics
University of Benin
writing

From the blog

all articles
Claude Code · AI Agents

AI-Native Delivery: Claude Code Agents, Linear and GitHub Actions

A short tutorial on turning Linear tickets into reviewed draft PRs with Claude Code agents on GitHub Actions, with payment code guarded so only humans can change it.

read
LangChain · Artificial Intelligence

Building Agentic AI Into Your Product Using LangChain (Node.js Edition)

Learn to how to build agentic AI into your Node.js application using LangChain, Google Gemini, and a clean, production‑ready architecture.

read
Node.js · Event-Driven

What Does Event-Driven Mean in Node.js?

Event‑driven architecture is a style where program flow is determined by events like user actions, timers, & network responses.

read
Javascript · ES2020

Modern JS Features You Should Use

JavaScript has evolved massively over the past few years. With recent development in JS, you can now write cleaner, safer, and more expressive code.

read
$ ./contact

Let's build something
worth shipping.

Open to AI-native full-stack & product engineering roles. Based in Berlin, working across timezones.

martin@martinoputa.comgithublinkedin