Description:
AI & Data Platform
The AI & Data Platform team powers intelligence across the Bullish ecosystem—from our institutional-grade cryptocurrency exchange to CoinDesk's media, data, and indices businesses. We build the infrastructure that transforms raw data into governed, trustworthy assets and deploy AI systems that meet the reliability standards our institutional clients expect.
We operate at the intersection of data engineering, semantic modeling, and applied AI—delivering solutions across the enterprise, spanning the full breadth of Bullish operations from trading floors to compliance.
As we expand our AI capabilities, we are building out a robust evaluation and governance layer—ensuring every AI system we deploy is observable, testable, and held to the same engineering standards as our trading infrastructure. This is a team where AI is treated as serious engineering, not experimentation.
What You'll Do
- AI Systems Architecture: Design and implement production AI systems with emphasis on reliability, observability, and continuous evaluation. Champion engineering practices that ensure AI outputs are consistent, evidence-based, and auditable.
- Conversational Analytics: Lead development of natural-language interfaces to business data, enabling stakeholders to query complex datasets through governed semantic layers.
- Agent Development: Architect multi-agent systems that coordinate across data sources, applying industry best practices for agent orchestration, tool use, and structured output generation.
- Evaluation & Trust: Build evaluation harnesses and testing frameworks that measure AI system quality—including groundedness, factual consistency, and output reliability—before deployment to production.
- Cross-Functional Collaboration: Work closely with product, trading, research, and media teams to translate complex requirements into scalable AI solutions with clear success metrics.
- Technical Leadership: Mentor engineers, establish coding standards, and drive architectural decisions that balance innovation velocity with production stability.
- Semantic Layer Partnership: Partner with data engineering to define and enforce semantic models that bridge raw data to AI-consumable formats, ensuring business logic is captured once and reused across applications.
What You'll Bring
AI Engineering Experience: 5+ years building production AI/ML systems, with demonstrated experience deploying LLM-based applications beyond proof-of-concept.
Agent & Orchestration Expertise: Hands-on experience with agent frameworks, tool-use patterns, and multi-step reasoning systems.
Experience With At Least 3 Of The Following
- LangChain / LangGraph (agent orchestration, stateful workflows)
- LlamaIndex (RAG pipelines, data indexing, query engines)
- Multi-agent frameworks (CrewAI, AutoGen/AG2, MetaGPT, OpenAI Agents SDK)
- Model Context Protocol (MCP) for tool connectivity
- DSPy (programmatic prompting, optimization)
- Vector databases (Qdrant, Pinecone, ChromaDB, Weaviate, pgvector)
- Structured output libraries (Instructor, Pydantic, Zod)
- LLM inference infrastructure (vLLM, Ollama, TensorRT-LLM)
- Cloud AI platforms (Vertex AI, Amazon Bedrock, Azure OpenAI)
- Evaluation & observability tools (LangSmith, Braintrust, Weights & Biases, Arize)