Senior Applied AI Engineer, Document Intelligence

Responsibilities: — Design and own the architecture of document processing pipelines including page classification, field extraction, candidate retrieval, claim matching, confidence gating, and hand-off to human review. — Decide the boundary between rules and models and define guarantees between pipeline stages. — Manage multiple pipelines for different document types, including emergence and consolidation of pipelines. — Evaluate new systems and run structured trials of candidate components, recommending what to keep. — Build evaluation processes for false positives, false negatives, and errors of omission to determine pipeline upgrades. — Ship and operate components in Python workers on the cloud, monitor costs and quotas, and diagnose reviewer feedback. — Collaborate with leadership on planning for accuracy, cost improvements, and overall product enhancements. Requirements: — 6+ years building ML or data-intensive systems, with at least 3 years owning a production pipeline. — Strong backend engineering and ideally cloud architecture understanding. — Hands-on experience with document understanding technologies such as OCR, layout models, field extraction from scanned PDFs, or intelligent document processing. — Experience with entity resolution or record linkage at scale using tools like Splink, Dedupe, or custom systems. — Experience designing evaluation processes from scratch including labeled ground-truth sets, metrics selection, and threshold defense to non-technical stakeholders. — Production experience with LLM APIs including structured output, tool calling, and cost or rate-limit management. — Strong skills in Python and SQL, comfortable with Postgres query plans and raw queries. — Scientific mindset towards testing pipelines and interpreting results. Nice to have: — Experience with Azure AI services or AWS/GCP equivalents. — Experience in financial document processing and regulated workflows handling PII. — Knowledge of probabilistic matching theory and active-learning labeling workflows. — Experience building human-review tooling or feedback loops for training and evaluation data. Benefits: — Small, tight-knit team with direct access to large datasets. — Weekly planning, daily standups, fully remote within US timezones. — Competitive compensation aligned with senior individual contributor role. — Monthly payments in USD. — Equal opportunity employer with merit-based hiring and non-discrimination policy.