4 hours ago
Remote, United StatesSenior
Base Salary
$165k - $200k/yr
Responsibilities
- Implement embeddings, vector storage, and retrieval across a large provenance-tracked evidence base.
- Integrate language models so generated text is tied to cited source records and test for citation failures.
- Design prompts and output schemas.
- Own model packaging, versioning, serving, and rollback.
- Instrument telemetry for retrieval and generation quality, recommendation and version attribution, overrides, abstentions, grounding failures, latency, throughput, and measurement events.
- Provide bounded model assistance for difficult narrative extraction, with every output tied to its source passage.
- Supply recorded rule context, including exact rule versions and ordered context, to every model-assisted step.
- Maintain a modular in-boundary serving path, self-hosted or managed, alongside the primary managed inference service.
Requirements
- At least 5 years of experience, including a production or near-production retrieval-augmented generation system built by the candidate.
- Detailed hands-on knowledge of retrieval design, vector stores, grounding tests, citation failures, and rollback practices.
- Strong Python skills and hands-on experience with embeddings and vector retrieval at scale.
- Ability to clearly distinguish shipped production code from prototypes or proposals.
- US citizenship is required.
- An active US Secret clearance is required to start.
Benefits
- Fully remote, results-based work environment with occasional travel to headquarters, customer sites, and vendor facilities.
- Competitive salary, bonus, and equity package.
- Employer-paid medical, dental, and vision insurance for the employee and family.
- Unlimited paid time off with manager approval.
- Flexible work environment where employees manage their workday.
- 14 weeks of fully paid parental leave.
Tech Stack
Categories
About DEFCON AI
DEFCON AI builds AI-driven simulation and decision-support software for logistics, mobility, and supply chain operations. Its platform uses mathematical optimization, machine learning, and data science to help organizations plan, assess risk, and keep missions and shipments moving during disruptions; current work includes delivering capabilities for the U.S. federal government. Founded in 2022 and headquartered in McLean, Virginia, the company is privately held and raised a seed round in 2024.