ML Engineer
Orcrist Technologies3 months ago
Remote, WorldwideMid Level / Senior
Responsibilities
- Package and deploy ASR, translation, OCR, NER, and summarization models using Triton and KServe on Kubernetes.
- Build evaluation pipelines using WER, BLEU, F1, latency, and cost metrics, and automate release gating.
- Operate streaming and batch inference using Kafka, Temporal, and backfill tooling.
- Monitor model drift and quality with Prometheus, Grafana, and Evidently while optimizing inference cost and performance.
- Collaborate with Research, product squads, and TypeScript teams on payload schemas, contracts, and human-in-the-loop feedback loops.
Requirements
- 4–8+ years of ML engineering or MLOps experience shipping models to production.
- Strong Python, PyTorch, and Transformers skills, with experience using Triton, KServe, or similar tools.
- Experience with Kubernetes, GPU workload operations, evaluation metrics, monitoring, and annotation workflows.
- Comfort with GitOps and CI/CD.
- Eligible to work in Germany; export-control screening is required for certain programs.
- Nice-to-have experience includes Temporal, Beam, Flink, Ray Serve, ONNX, TensorRT, WhisperX, DeepStream, GStreamer, or vector search integrations.
- German language proficiency at B1+ and familiarity with defense or public safety datasets are advantageous.
Benefits
- Remote-first role in Germany with regular Berlin meetups.
- 30 days of vacation.
- Equipment and learning budget.
Tech Stack
Categories
About Orcrist Technologies
Orcrist Technologies offers pioneering AI and data analytics solutions in the private and public sectors, turning sensors into strategy. We are a Berlin-based data defense technology company building AI-powered software for real-time situational awareness and sensor fusion. Our mission is to give decision-makers the clarity they need—when it matters most. Designed for interoperability, speed, and modularity, Orcrist enables NATO and European partners to operate with information dominance in fast-moving, high-threat scenarios.