About The Role
The role sits at the intersection of LLM engineering and production systems - building the GenAI applications, retrieval pipelines, and evaluation infrastructure that power real products used every day. This is not a research position; it is hands-on engineering work where hallucination rates, latency budgets, and cost per token are all metrics that matter.
The team ships fast and iterates hard. Expect to own end-to-end AI features - from architecture and prototyping through deployment, monitoring, and continuous evaluation - working alongside applied scientists, platform engineers, and product teams.
Key Responsibilities
Design and build production RAG pipelines using LangChain, LlamaIndex, or custom architectures, including chunking strategies, reranking, and hybrid retrieval
Integrate and optimize vector databases (Pinecone, Weaviate, pgvector, Milvus) for semantic search workloads at scale
Develop systematic evaluation frameworks for LLM applications: benchmark suites, LLM-as-judge pipelines, human-in-the-loop review, and regression testing against production traffic
Fine-tune and adapt models using parameter-efficient techniques (LoRA, QLoRA) and manage the full training-to-deployment workflow
Build agentic workflows and tool-use chains with structured outputs, function calling, and robust error handling
Deploy LLM services behind APIs with attention to latency, streaming, caching, rate limiting, and cost optimization
Instrument LLM systems for observability - tracing, drift detection, safety guardrails - and drive improvements from monitoring data
What We Are Looking For
3–6 years of software or ML engineering experience, including at least 1–2 years building LLM/GenAI applications in production
Strong Python engineering skills; comfort with async programming, REST APIs, and building reliable services
Deep hands-on experience with at least one LLM orchestration framework (LangChain, LlamaIndex) and modern model APIs (OpenAI, Anthropic, or open-source models via vLLM/TGI)
Practical understanding of embeddings, vector search, chunking strategies, and retrieval quality tradeoffs
Experience with cloud infrastructure (AWS, GCP, or Azure), Docker, and CI/CD practices for ML services
MS in Computer Science or related field, or equivalent practical experience shipping production AI systems
Bonus: Experience fine-tuning open-weight models (Llama, Mistral), prompt injection defenses and LLM security, Kubernetes-based model serving, or contributions to open-source LLM tooling