Job
Description
Key Responsibilities: Design and develop a modular, scalable AI platform to serve foundation model and RAG-based applications. Build pipelines for embedding generation , document chunking , and indexing . Develop integrations with vector databases like Pinecone , Weaviate , Chroma , or FAISS . Orchestrate LLM flows using tools like LangChain , LlamaIndex , and OpenAI APIs . Implement RAG architectures to combine generative models with structured and unstructured knowledge sources. Create robust APIs and developer tools for easy adoption of AI models across teams. Build observability and monitoring into AI workflows for performance, cost, and output quality. Collaborate with DevOps, Data Engineering, and Product to align platform capabilities with business use cases. Core Skill Set: Strong experience in Python, with deep familiarity in ML/AI frameworks (PyTorch, Hugging Face, TensorFlow). Experience building LLM applications , particularly using LangChain , LlamaIndex , and OpenAI or Anthropic APIs . Practical understanding of vector search , semantic retrieval , and embedding models . Familiarity with AI platform tools (e.g., MLflow, Kubernetes, Airflow, Prefect, Ray Serve). Hands-on with cloud infrastructure (AWS, GCP, Azure) and containerization (Docker, Kubernetes). Solid grasp of RAG architecture design , prompt engineering , and model evaluation . Understanding of MLOps, CI/CD, and data pipelines in production environments. Preferred Qualifications: Experience designing and scaling internal ML/AI platforms or LLMOps tools. Experience with fine-tuning LLMs or customizing embeddings for domain-specific applications. Contributions to open-source AI platform components. Knowledge of data privacy, governance, and responsible AI practices. What You’ll Get: A high-impact role building the core AI infrastructure of our company. Flexible work environment and competitive compensation. Access to cutting-edge foundation models and tooling. Opportunity to shape the future of applied AI within a fast-moving team. Show more Show less
