Senior AI/ML Engineer
Company: SimpliGov LLC
Location: Baltimore, Maryland, United States (Remote)
Salary: $185k - $215k per year
Type: Full-time
Remote: Yes
Posted: 2026-08-12
About this role
Role Overview:
SimpliAI is our AI product suite (Ask, Forms, Chat, and Build) moving from private preview toward general availability for state and local government customers. You will design, build, and ship the AI systems behind it: agentic workflows, retrieval, evaluation, model routing, and the guardrails that make AI trustworthy enough for government.
We are explicit about what we are hiring for. AI capability is table stakes here; we will verify it, not celebrate it. What we actually screen for is engineering judgment: the ability to take an ambiguous problem, make defensible design decisions, and hand back work you would stake your name on. Our non-negotiable standard is that you own and can explain every line you ship, no matter what tool produced it.
Responsibilities:
- Design, build, and ship production AI features across the SimpliAI suite: agentic workflows, retrieval-augmented generation, structured extraction, and form and workflow intelligence
- Build evaluation before features: golden datasets, calibrated LLM-as-judge scoring, and regression evals wired into our observability stack; "it demos well" is not a bar we recognize
- Own the quality, safety, latency, and cost of what you ship, including per-workload model routing and unit economics
- Operate inside our compliance boundary: self-hosted observability, GovCloud inference paths (including Amazon Bedrock), and disciplined data retention
- Work in the open through our Plan-and-Review cadence: written designs, explicit assumptions, uncertainty surfaced early
- Raise the AI floor of the whole team: reusable patterns, reviews, and shared components, not private magic
Qualifications:
- 5+ years of software engineering with production ownership, including 2+ years shipping LLM or ML-backed features to real users
- Demonstrated evaluation discipline: you can show how you measured a system, not just that you built it
- Hands-on depth with modern LLM stacks: agentic patterns, RAG, ...