xelys jobs xelys jobs

Data & ML Engineer

DEFCON AI

full-remoteseniorpermanentdatabackend Full remote 98 days ago via WTTJ

See how well this job matches your profile

Sign up to get an AI match score and generate a tailored application in seconds.

Get your match score

Tags

PythonSQLProbabilistic Record MatchingRecord LinkageMaster Data ManagementIdentity ResolutionGraph Data ModelingModel CalibrationRetrieval-Augmented Generation (RAG)Responsible AI

About the role

Role Overview

Build the data and ML/model layer behind an AI-enabled decision-support system. You will focus on new capabilities such as secure ingestion, probabilistic record matching, relevance/priority modeling, and user-facing explanations that remain traceable to underlying sources.

Key Missions

  • Design and implement data models and record matching (including probabilistic matching).
  • Develop relevance and priority models over large, imperfect record sets, including calibration and threshold design.
  • Build secure pipelines for ingestion, transformation, validation, and publication of structured, semi-structured, and unstructured data.

Responsibilities

  • Deliver operational systems meant for sustained use (not just exploratory analysis).
  • Ensure every system output is grounded and traceable back to source data.
  • Create and validate retrieval-augmented generation (RAG) in production, including prompt and output-schema design and testing for grounding.
  • Support production operation of models (self-hosted/open-weight) including fine-tuning, adapters, or custom embeddings.
  • Follow responsible AI practices (documentation, fairness testing, and model monitoring).

Requirements

  • US Citizenship required; elevated personnel security requirements apply.
  • Active US Secret clearance (Top Secret required per posting) and ability to operate from the start in a controlled government cloud environment with favorable investigation and CAC eligibility.
  • 5+ years in data engineering, data architecture, applied/ML engineering, or production analytics engineering.
  • Strong Python and SQL, with demonstrated experience working with large, imperfect operational data.
  • Ability to explain a technical decision to stakeholders who must defend it without needing internal details.
  • Direct experience with probabilistic record linkage / identity matching on inconsistent identity data (names, dates, addresses, identifiers), including knowledge of failure modes.
  • Experience relevant to record linkage/MDM/identity management and graph data modeling/graph algorithms in production.

Nice-to-haves / Additional Details

  • Willingness to travel up to 25% to customer sites, DEFCON AI HQ, and vendor facilities.
  • Routine use of AI-assisted development, with judgment about when outputs require verification.
  • Experience with:
    • PostgreSQL and pgvector (or comparable), plus graph algorithms in production
    • Model calibration and threshold design, cost-sensitive learning
    • ML frameworks: scikit-learn, XGBoost, PyTorch
    • RAG: prompt/output schema design; grounding tests
    • Pipelines: AWS Glue, Airflow, dbt, Spark, Kafka, NiFi
    • AWS/federal DevSecOps environments and accreditation/deployment (RMF, ATO/DoW cloud environments)
    • Responsible AI: NIST AI RMF or comparable practice
  • Domain: sensitive federal/defense data and privacy/comparable handling constraints.

Location / Travel

  • Full remote
  • Travel up to 25% as required.

About DEFCON AI

DEFCON AI develops an AI-enabled decision-support system that relies on robust data ingestion, retrieval, and explainable modeling. The work focuses on building a secure, traceable data and model layer for users, including probabilistic record matching and relevance scoring.

Scraped 8/16/2026