SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: EUR 6,500 - 9,000 / annual
Cast AI's Kimchi is an AI platform enabling companies to run large language models on their own Kubernetes clusters and managed variants. The platform provides multi-model inference (MiniMax, Kimi, GLM-5, Nemotron, DeepSeek) with intelligent routing, an OpenAI-compatible API, and flexible deployment options from Cast AI's GPUs to customer VPCs.
The Harness team is building the scaffolding that transforms raw models into autonomous agents capable of planning, executing, recovering, and completing complex work. Unlike OpenAI and Anthropic's harnesses, Cast AI's harness is purpose-built for cost-conscious, long-horizon autonomy running on inference infrastructure controlled end-to-end. The team believes that a good model with a great harness outperforms a great model with a poor harness—and the harness gap is where significant value lies.
As a Senior Software Engineer on the Harness team, you will own features end-to-end: design, implementation, testing, and production rollout. Most projects ship in 1–4 weeks. You'll work directly with product and other engineering teams on problems without textbook solutions, minimizing process friction to enable rapid shipping.
Key responsibilities include:
- Architecting planner/executor/evaluator pipelines: planning with a reasoning model, execution with a fast model, evaluation with a third model, without self-verification loops.
- Managing agent memory and context: state persistence across sessions, context compaction, and tool-call offloading.
- Owning the harness surface: TUI, MCP integrations, and telemetry.
- Working directly with users—including colleagues—to identify, understand, and fix pain points.
You'll have high agency across multiple parts of the tech stack and will drive complex projects to completion in a startup environment that values adaptability and proactivity.
REQUIREMENTS:
- Production experience with Go or TypeScript strongly preferred; candidates without either must demonstrate strong systems programming skills in a comparable language.
- Strong debugging, optimization, and performance-tuning skills, including query profiling, index design, and database performance tuning beyond ORM usage.
- Hands-on experience with cloud platforms (AWS, GCP, or Azure) and Kubernetes is a strong plus.
- Experience with observability tooling (Prometheus, Grafana, OpenTelemetry), CI/CD, and DevOps practices.
- Startup mindset: adaptable, proactive, and comfortable with ambiguity.
- Strong English skills, both verbal and written.
- Demonstrated ability to personally drive a complex project end-to-end.
- Experience working with harnesses and creating custom AI workflows is a strong plus.
