Internal ai frameworks, sdks, and shared libraries
Standardized patterns for llms, embeddings, agents
Reusable components for prompt management, evaluation, safety
Design and build internal AI frameworks, SDKs, and shared libraries to enable teams to integrate AI features with minimal friction
Job Summary
Design and build internal AI frameworks, SDKs, and shared libraries to enable teams to integrate AI features with minimal friction.
Set up standardized patterns for using LLMs, embeddings, agents, and workflows, building reusable components for prompt management, evaluation, observability, and safety.
Evangelise AI internally through documentation, examples, and hands-on guidance, rapidly prototyping AI-powered features and turning them into reusable building blocks.
Matching Summary
Design and build internal AI frameworks, SDKs, and shared libraries to enable teams to integrate AI features with minimal friction.
Skills & Requirements
Must-have
Internal AI frameworks, SDKs, and shared libraries
Standardized patterns for LLMs, embeddings, agents
Reusable components for prompt management, evaluation, safety
Best practices for AI usage, cost control, and reliability
Hands-on experience with LLM APIs
Experience with RAG pipelines, embeddings, vector databases
Excellent Python skills
Experience building frameworks, platforms, or developer tooling
Nice-to-have
Experience with agent-based systems and tool calling
Experience setting up guardrails, moderation, and safety checks
Familiarity with Docker, cloud infra, and CI/CD
Prior work in platform teams or enablement roles
Key Requirements
Shipped at least one LLM-powered feature to production
Comfortable with embeddings, fine-tuning, vector search, tokenization, and evaluation methodologies
Familiar with data lakes/warehouses, feature stores, and streaming/batch ETL; cloud (AWS/GCP/Azure)
Strong understanding of APIs, services, and system design
Writes clean, maintainable, production-grade code
Bias for shipping and iteration
Can translate AI complexity into simple abstractions
Comfortable with ambiguity and rapid change
Thinks in terms of leverage and reuse, not one-off solutions