Applied AI/ML Engineer
AI summary of the role
Senior Applied AI/ML Engineer to build the core AI behind Confido, an AI infrastructure platform for CPG brands.
What you’ll do
- Build LLM systems that turn complex, inconsistent financial documents into clean, typed, validated data
- Design agentic workflows that retrieve and reason over fragmented enterprise systems, with eval harnesses and guardrails for production
- Build and improve demand forecasting across thousands of intermittent series using classical time-series and gradient-boosted models
- Wrangle messy real-world data: baselining, outlier/anomaly detection, entity reconciliation, and ingestion across heterogeneous sources
What you’ll bring
- 3+ years of applied AI/ML with systems taken to production
- Depth in at least one of: time-series/forecasting, LLM/agentic systems, or large-scale messy-data engineering; working fluency across several
- Experience building evaluation and monitoring, avoiding leakage and train/serve skew
- Strong product sense and ability to turn AI capability into real outcomes
Technologies
LLM · agentic workflows · RAG · time-series forecasting · Prophet · ARIMA · gradient-boosted models · document understanding · information extraction · eval harnesses
Source and classification
Internal deployment & tooling · Evidence for this classification:
Confido is the AI infrastructure powering modern CPG — the platform that 200+ brands like OLIPOP, Simple Mills, Dr. Squatch, and Tropicana use to run everything from deductions to production planning. Finance, accounting, sales, and operations, unified in one system for the first time. We're growing 5x year over year with a small team in New York City; the people who join now will shape the product, the culture, and the company itself. If you want your work on shelves everywhere — and outsized ownership while you build — we'd love to meet you. The Role Build the core AI behind Confido — and get your hands on an unusually rich, messy dataset: hundreds of thousands of documents across dozens of sources and hundreds of layouts and contexts, much of it beyond what any system reads well today. As a Senior Applied AI / ML Engineer, you'll own AI problems end to end — research,
More from the job description
Confido is the AI infrastructure powering modern CPG — the platform that 200+ brands like OLIPOP, Simple Mills, Dr. Squatch, and Tropicana use to run everything from deductions to production planning. Finance, accounting, sales, and operations, unified in one system for the first time. We're growing 5x year over year with a small team in New York City; the people who join now will shape the product, the culture, and the company itself. If you want your work on shelves everywhere — and outsized ownership while you build — we'd love to meet you. The Role Build the core AI behind Confido — and get your hands on an unusually rich, messy dataset: hundreds of thousands of documents across dozens of sources and hundreds of layouts and contexts, much of it beyond what any system reads well today. As a Senior Applied AI / ML Engineer, you'll own AI problems end to end — research, prototyping, production, and the evaluation that keeps probabilistic systems reliable — across document understanding, forecasting, and agentic workflows. Location: New York, NY (Relocation supported) What you'll do Build LLM systems that turn complex, inconsistent financial documents into clean, typed, validated data — the backbone the rest of the platform runs on Design agentic workflows that retrieve and reason over fragmented enterprise systems, with the eval harnesses and guardrails to run them in [... source excerpt omitted ...] eval datasets, benchmarks, and monitoring for systems with no single "correct" answer What we're looking for Required 3+ years of applied AI / ML, with systems you've taken to production (not just prototypes) Depth in at least one, and working fluency across several, of: time-series / forecasting and statistical modeling; LLM and agentic systems; large-scale messy-data engineering You build evaluation and monitoring as a matter of course — and know which metric to trust, and how to avoid leakage and train/serve skew Strong product sense: you turn AI capability into real outcomes, and know when a simpler approach wins Masters degree in STEM Nice to have Demand forecasting
Employer postings · Data from · Sources