Engineer deterministic RAG pipelines, cognitive vector retrieval engines, and production LLM orchestration systems for institutional clients.
“High-stakes software requires engineers who take extreme ownership of failure domains, developer velocity, and long-term maintainability.”
Enterprise clients cannot afford probabilistic hallucinations in compliance, legal, and financial workflows. This role exists to bridge raw LLM intelligence with deterministic verification pipelines, hybrid vector search architectures, and automated evaluation harnesses that deliver guaranteed factual accuracy.
Architect resilient pipelines that sustain high throughput under volatile load without performance degradation.
Eliminate race conditions, unhandled failure modes, and architectural ambiguity with rigorous typing and automated verification.
Drive technical direction, author Technical Decision Records (ADRs), and guide squad standards with total autonomy.
You will not be micromanaged through tickets. You are trusted to operate as the technical authority for these core domains:
Design hybrid dense-sparse indexing and reciprocal rank fusion across millions of documents.
Enforce automated citation verification to eliminate factual hallucinations in production.
Architect multi-tenant pgvector and Qdrant clusters with strict cryptographic separation.
Deploy continuous LLM benchmarks measuring recall, context precision, and latency SLAs.
Tangible distributed software, deterministic AI infrastructure, and design systems deployed into global production.
Reciprocal Rank Fusion pipeline merging vector distances with BM25 inverted indexes.
Sub-second citation hashing engine checking generated statements against ground truth.
LangGraph stateful workflow system coordinating specialized document synthesis agents.
We operate as tightly integrated product squads with transparent communication and minimal meeting fatigue.
Autonomous deep-work blocks with asynchronous experiment logs and RFC reviews.
Clear, transparent milestones so you always know where you stand and what is expected of you.
Evaluate current vector retrieval precision and deploy your first evaluation benchmark.
Ship hybrid retrieval architecture reducing retrieval latency below 45ms P95.
Deploy deterministic citation verification across 500,000+ live enterprise queries.
We value deep craftsmanship, clear systems thinking, and production experience over credentialism.
A fast, respectful, and practical 4-step interview journey designed by engineers for engineers. No algorithmic leetcode puzzles or artificial trick questions.
A casual conversation with our engineering leadership to discuss your background, career goals, and ensure mutual cultural alignment.
No whiteboard leetcode puzzles or artificial trick questions. We dive into a real-world system architecture challenge relevant to our daily client work.
Review a real pull request or open-source contribution together with the peers you would be working alongside daily.
Transparent, market-leading compensation offer with clear leveling, equity options where applicable, and immediate start onboarding.
Our Commitment to Candidates: We respond to every application within 5 business days, provide specific technical feedback after architecture reviews, and make offers within 48 hours of final rounds.
| Department | AI Systems |
|---|---|
| Seniority Level | Staff |
| Location & Model | Remote (Worldwide) — Remote-First (Async Cadence) |
| Employment Type | Full-time |
| Compensation Benchmark | $160,000 – $210,000 USD + Equity Options (Based on technical contribution, not geography) |
| Primary Tech Stack | PythonPyTorchpgvectorLangGraphFastAPIDockerAWS Bedrock |
Reviewed directly by senior engineering leads. We respond to every application within 5 business days.
