Fabrion
Data Engineer (Founding Team)
11mo ago
USASenioretldata pipelinesknowledge graphsstreamingsemantic enrichment
Build and maintain scalable data pipelines and connector frameworks for a multi-tenant AI-native data platform.
Responsibilities
- We’re building a multi-tenant, AI-native platform where enterprise data becomes actionable through semantic enrichment, intelligent agents, and governed interoperability. At the heart of this architecture lies our Data Fabric — an intelligent, governed layer that turns fragmented and siloed data into a connected ontology ready for model training, vector search, and insight-to-action workflows.
- We're looking for engineers who enjoy hard data problems at scale : messy unstructured data, schema drift, multi-source joins, security models, and AI-ready semantic enrichment. You’ll build the backend systems, data pipelines, connector frameworks, and graph-based knowledge models that fuel agentic applications.
- If you've worked on streaming unstructured pipelines, built connectors into ugly legacy systems, or mapped knowledge graphs that scale — this role will feel like home.
- Build highly reliable, scalable data ingestion and transformation pipelines across structured, semi-structured, and unstructured data sources
- Develop and maintain a connector framework for ingesting from enterprise systems (ERPs, PLMs, CRMs, legacy data stores, email, Excel, docs, etc.)
- Design and maintain the data fabric layer — including a knowledge graph (Neo4j or Puppygraph) enriched with ontologies, metadata, and relationships
- Normalize and vectorize data for downstream AI/LLM workflows — enabling retrieval-augmented generation (RAG), summarization, and alerting
- Create and manage data contracts, access layers, lineage, and governance mechanisms
- Build and expose secure APIs for downstream services, agents, and users to query enriched semantic data
- Collaborate with ML/LLM teams to feed high-quality enterprise data into model training and tuning pipelines
Nice to have
- Prior work with vector DBs (e.g. Weaviate, Qdrant, Pinecone) and embedding pipelines
- Experience building or contributing to enterprise connector ecosystems
- Knowledge of ontology versioning , graph diffing , or semantic schema alignment
- Familiarity with data fabric patterns (e.g. Palantir Ontology, Linked Data, W3C standards)
- Familiar with fine-tuning LLMs or enabling RAG pipelines using enterprise knowledge
- Experience enforcing data access policy with tools like OPA , Keycloak , Snowflake row-level security
Other
- Location: San Francisco Bay Area
- Type: Full-Time
- Compensation: Competitive salary + early-stage equity
- Backed by 8VC, we're building a world-class team to tackle one of the industry’s most critical infrastructure problems.
- 5+ years building large-scale data infrastructure in production environments
- Deep experience with ingestion frameworks (Kafka, Airbyte, Meltano, Fivetran) and data pipeline orchestration (Airflow, Dagster, Prefect)
- Comfortable processing unstructured data formats: PDFs, Excel, emails, logs, CSVs, web APIs
- Experience working with columnar stores, object storage, and lakehouse formats (Iceberg, Delta, Parquet)
- Strong background in knowledge graphs or semantic modeling (e.g. Neo4j, RDF, Gremlin, Puppygraph)
- Familiarity with GraphQL, RESTful APIs, and designing developer-friendly data access layers
- Experience implementing data governance : RBAC, ABAC, data contracts, lineage, data quality checks
- You’re a system thinker: you want to model the real world, not just process it
- Comfortable navigating ambiguous data models and building from scratch
- Passionate about enabling AI systems with real-world, messy enterprise data
- Pragmatic about scalability, observability, and schema evolution
- Value autonomy, high trust, and meaningful ownership over infrastructure
- Agents are only as smart as the data they operate on. This role builds the foundation — the semantic, governed, connected substrate — that makes autonomous decision-making and agent action possible. From factory ERP records to geopolitical news alerts, the data fabric unifies it all.
- If you're excited to tame complexity, unify chaos, and power intelligent systems with trusted data — we’d love to hear from you.