Founding Engineer (Backend/Platform & LLM)
NEW Gruppe
Berlin, Germany
22 days ago
Role details
Contract type
Permanent contract Employment type
Full-time (> 32 hours) Working hours
Regular working hours Languages
English Experience level
SeniorJob location
Berlin, Germany
Tech stack
Continuous Integration
Python
PostgreSQL
OAuth
OpenID
Role-Based Access Control
Security Assertion Markup Language (SAML)
SQLAlchemy
Systems Integration
Okta
Large Language Models
Backend
FastAPI
Playwright
Kafka
Amazon Web Services (AWS)
Docker
Job description
- Build and evolve our Python/FastAPI backend, typed data contracts, Postgres models, and deterministic code-gen pipeline (structured outputs, snapshot tests).
- Implement and tune LLM orchestration pipelines: RAG flows, agent workflows, evaluation harnesses, tracing, fallback logic, and cost/latency controls.
- Own CI/CD, IaC, migrations, SLOs/SLIs, observability, and the reliability of production systems.
- Integrate secure multi-tenant SSO (OIDC/SAML) with RBAC and audit trails for early healthcare pilots.
- Collaborate closely with the founders, influence product direction, and establish engineering standards.
Requirements
- 8-12+ years as a senior/principal backend or platform engineer with end-to-end ownership of production services.
- Strong experience with Python 3.10+, FastAPI, Postgres, SQLAlchemy/Alembic, Docker, CI/CD.
- Hands-on with LLM systems in production (RAG/agents, evaluation, tracing, performance/cost optimization).
- Solid operational mindset: SLOs, incidents, rollbacks, migrations, and running systems that "just work."
- Experience integrating Auth0/Keycloak/Okta and OAuth2/OIDC/SAML for multi-tenant environments.
- Nice addition: Playwright/pywinauto, vector DBs (pgvector/Qdrant), Temporal/Kafka/SQS, HIPAA/SOC2 exposure.
About the company
At New Objectives, we are building Europe's first secure, auditable, and EU-compliant automation infrastructure for legacy systems. We're taking a practical approach to improving how engineers work with AI systems. Combining LLMs with familiar software engineering methods to make it easier for people to build, modify and maintain software automations. Our platform lowers the effort required to turn natural-language instructions into deterministic, testable code that can interface with and automate legacy systems.