DelphiHire With Rapha

AI Engineer

San Francisco, CA (Onsite)full-time

Benefits

Competitive SalaryEarly Employee EquityHealthcare CoverageLearning StipendMacbook ProStudio DisplayUnlimited Book StipendGreat Office VibesAutonomyProduct OwnershipBuilding The Future!Glory! Glory! Glory!

About company

Delphi is a digital cloning platform that makes the knowledge and experiences in people’s minds available to everyone, enabling users to get personalized advice from those who have already achieved their version of success.

On the demand side, imagine a more direct & personalized Masterclass. What if you could get startup advice from Paul Graham? Negotiation strategies from Chris Voss?

Personalized to your specific circumstance? Well, now you can. On the supply side, Delphi allows people to scale their time & influence, so that a wider audience can access their knowledge & insights, in a more personalized way.

We are backed by the best from Founders Fund, Lux Capital, XFund & MVP Ventures, and Balaji Srinivasan

About the role

Our "Clone Brain" architecture represents a quantum leap in digital mind modeling, powered by sophisticated knowledge graphs that capture not just information, but the intricate web of associations, reasoning patterns, and conceptual frameworks that make each mind unique. While traditional RAG systems treat knowledge as a flat collection of documents, we're building rich, hierarchical representations that mirror how human experts actually think and reason about their domains.

Enter the Knowledge Graph Engineer. You'll architect and implement the neural-symbolic backbone of our Clone Brain system, developing graph structures that can capture both explicit knowledge and implicit reasoning patterns. If you're passionate about knowledge representation, graph neural networks, and pushing the boundaries of how machines can model human thought—this role is crafted for you.

What You Will Work On

  1. Graph Architecture & Design

    • Design and implement scalable knowledge graph architectures that can represent both declarative knowledge and procedural reasoning patterns

    • Develop novel graph embedding techniques that capture the nuanced relationships between concepts, experiences, and decision-making frameworks

    • Create efficient indexing and retrieval mechanisms for real-time graph querying

  2. RAG System Innovation

    • Build next-generation retrieval systems that combine traditional RAG with graph-based reasoning

    • Implement hybrid architectures that seamlessly integrate symbolic graph operations with neural retrieval

    • Develop evaluation frameworks to measure and optimize retrieval quality across different types of queries

  3. Knowledge Integration & Maintenance

    • Design systems for automated knowledge graph construction and maintenance from various data sources

    • Implement validation and consistency checking mechanisms to ensure graph quality

    • Create tools for experts to review and refine their knowledge representations

  4. Infrastructure & Scaling

    • Build infrastructure for distributed graph operations across thousands of clone instances

    • Optimize graph operations for low-latency retrieval in production environments

    • Develop monitoring and debugging tools for graph-based reasoning systems

Preferred Abilities

  • Graph Expertise: Deep experience with graph databases, knowledge graphs, and graph neural networks

  • RAG Systems: Strong background in retrieval-augmented generation, including both traditional and graph-based approaches

  • Neural-Symbolic Systems: Experience combining symbolic reasoning with neural networks

  • Distributed Systems: Ability to design and implement distributed graph processing systems

  • Python & Graph Tools: Proficiency with Python and graph processing frameworks (Neo4j, DGL, PyG, etc.)

  • Research Translation: Proven ability to translate academic research into production-ready systems

  • First Principles Thinking: Capacity to reason about and solve novel technical challenges without established playbooks

Why You Might Like This Role

  • Opportunity to work on fundamental AI challenges at the intersection of knowledge representation and cognitive modeling

  • High ownership over core technical architecture that powers thousands of digital minds

  • Chance to define new standards for how machines represent and reason with knowledge

  • Collaboration with a team pushing the boundaries of digital consciousness

  • Real product impact with paying customers while working on bleeding-edge technology

Why You Might Not Like This Role

  • Bleeding Edge Territory We're developing systems that have never existed before—there's no Stack Overflow answer for many of our challenges.

  • High Technical Uncertainty Many of our problems require novel solutions that combine multiple emerging technologies in untested ways.

  • Product-Research Balance We need to balance ambitious research goals with practical product requirements and customer needs.

  • Fully On-Site We believe in-person collaboration drives better ideas. If you're looking for remote, this might not be for you.