Required skills

Good to have

• Federated learning, differential privacy, or other privacy preserving training methods.
• Custom CUDA or Triton kernels.
• Model distillation and model merging.
• Guardrails, red teaming, and jailbreak testing.
• Multimodal models: vision, speech, or document understanding.
• Open source contributions to any part of this stack.
• Experience with on-premise or air-gapped deployment.

The AI Full Stack Developer role combines language-model training and alignment with agent workflows, inference, and backend services. The required stack spans Python, ML frameworks, evaluation methods, data retrieval, and production operations; several privacy, kernel, and multimodal topics are listed as additional advantages.

Role at a glance

Skills mentioned

What you will likely work on

• Train and fine-tune models using supervised methods and parameter-efficient techniques.
• Apply at least one listed alignment approach and assess its measured results.
• Build agent flows using tool calling, context management, and evaluation.
• Serve models and improve inference through quantization and latency tuning.
• Develop Python backend services and connect them to data stores.
• Create retrieval workflows using embeddings, vector databases, and RAG techniques.
• Evaluate model behavior with benchmarks, custom sets, LLM-based judging, and regression tests.
• Support deployment workflows using containers, CI/CD, Linux, and GPU resources.

Good fit signals

• Expert-level Python and working proficiency in TypeScript or JavaScript match the stated language requirements.
• Hands-on work with at least one of RLHF, DPO, GRPO, or RLVR, including measured results, is requested.
• Experience across model training, agent systems, inference, and backend development aligns with the full-stack scope.
• Familiarity with evaluation practices and GPU planning is relevant to the stated stack.
• Privacy-preserving training, custom kernels, multimodal systems, and air-gapped deployment are helpful additional experience, not core requirements.