RAG & AI Agent Engineer (Remote) | AI Startup

About the job

Join us to take RAG and AI Agents from 1→10 and then to scale. This is a founding role. In addition to salary, we offer meaningful equity to reflect the ownership and long-term impact of this position. You'll own the RAG/agent roadmap, set engineering standards, and influence product direction and early culture.

Our Vision

Our goal is simple: make world-class analysis and creation as repeatable as running code, and put it in the hands of every team.

We believe the future of knowledge work is AI-native: teams won't sift through files—they'll collaborate with agents that understand context, reason over organizational memory, and produce publish-ready work. With Weavin, we're building that future: a living knowledge fabric where RAG + AI Agents turn scattered, multimodal inputs into structured, reusable assets measured by quality, speed, and cost.

Where we work & how we work:
We're a small, high-trust crew operating from Sydney, Shenzhen, and Hangzhou. We ship weekly, obsess over users, and favor ownership, kindness, and clear thinking.

What you'll do

• RAG architecture & implementation:
  parsing/chunking, embeddings, indexing, search, re-ranking, generation & grounded citation. Support multilingual and multimodal content.
• Search quality & relevance:
  hybrid search (vector + BM25/keywords), query rewrite (Rewrite/HyDE/intent classification), Cross-Encoder or third-party ReRank.
• Agent orchestration (LangGraph):
  stateful, multi-tool, multi-node agents with checkpoints, pause/resume, human-in-the-loop; integrate search, file processing, and structured outputs.
• Evaluation & observability:
  offline/online evals (accuracy, coverage, citation hit rate, latency, cost), tracing/logging/alerts, and A/B tests.
• Platform engineering:
  stable API/SDKs, multi-tenant RBAC isolation, caching, rate limiting, retries/fallbacks, graceful degradation, and model rollbacks; compliance & audits.

What you'll bring

• Strong coding in Python (preferred) or TypeScript; clean code, tests, CI/CD.
• Hands-on LangGraph (State/Channels/Node/Edge/Subgraph; events & checkpoints using SQLite/Redis/custom).
• RAG depth: chunking strategies, metadata indexing, hybrid search & ReRank.
• Vector DB experience: FAISS / Milvus / pgvector / Elastic (any).
• Prompt/Tool/Function calling; structured outputs & citation alignment.
• Reliability mindset: large documents and frequent updates, backfills; speed/timeouts/retries; multi-tenant isolation.
• Collaboration: turn complex workflows into clean interfaces & reusable templates with a focus on maintainability and UX consistency.

Nice-to-haves

• LangSmith or OpenTelemetry for tracing & dashboards.
• Multimodal RAG (tables/images/audio/video) or structured extraction/knowledge graphs.
• Performance & cost tuning (caching, prompt compression, tiered recall, multi-level storage).
• FastAPI, messaging (Kafka/Redis Streams), containers & cloud (Docker/K8s/major clouds).
• Open-source contributions, tech blogging, or talks.

Why us

• You can work remote-first with flexible arrangements, and we are happy to support three or four days a week if that suits your life.
• You will join in a founding capacity with meaningful equity because your decisions will shape the product, the stack, and the culture.
• You will be part of a people-first, low-ego environment with supportive reviews, clear communication, and an inclusive culture where you can bring your whole self.
• You will collaborate with a friendly, global crew across Sydney, Shenzhen, and Hangzhou who share diverse perspectives and a single standard for great work.

Apply

Email your GitHub/portfolio + CV to with subject "RAG & Agent Engineer".