> Markdown version of [/jobs/ext/3538196-senior-ai-ml-engineer](https://www.wearedevelopers.com/jobs/ext/3538196-senior-ai-ml-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Senior AI/ML Engineer

- **Company:** Colvin Run Networks, Inc.
- **Location:** United States
- **Experience:** Expert
- **Salary:** $160,000.0 - $210,000.0
- **Contract:** Permanent contract
- **Skills:** LangGraph Framework, Java (Programming Language), Artificial Intelligence, Amazon Web Services, Microsoft Azure, Continuous Integration, Data Infrastructure, Python (Programming Language), Machine Learning, Software Architecture, Power BI, Tensorflow, SQL Databases, Web Services, Pytorch, LangChain, Retrieval-Augmented Generation, Llamaindex, Keras, Git, Scikit Learn, Kubernetes, Information Technology, Machine Learning Operations, Drift Detection, GPT, Software Version Control, Docker, Databricks
- **Published:** October 2, 2026
- **Apply:** https://www.clearancejobs.com/jobs/9213693/senior-aiml-engineer

## About the Role

* Bachelor's degree in computer science, engineering, mathematics, data science, or a related field.
 * 8 or more years of relevant experience, including at least 3 years working with AI technologies.
 * Hands-on production experience designing, training, and deploying machine learning models.
 * Strong Python programming, plus working knowledge of Java or R.
 * Experience with machine learning frameworks such as PyTorch, TensorFlow, or Keras, and libraries such as scikit-learn.
 * Working knowledge of statistics, data modeling, and software architecture as they apply to ML systems.
 * Experience with Git, CI/CD pipelines, containers (Docker or Kubernetes), and SQL.
 * Must be a U.S. citizen with the capability of obtaining a Top Secret security clearance.

Desired Experience

 * Use of LLM frameworks such as LangChain, LangGraph, LlamaIndex, or MCP.
 * MLOps practices, including model versioning, monitoring, and automated retraining.
 * Delivery under federal authorization requirements such as ATO, RMF, or FISMA.
 * Work with Databricks or a comparable data platform, and with AWS or Azure.
 * Use of a business intelligence or dashboard tool such as Power BI.
 * Explaining a model's limits to people who do not build models.

## Description

Colvin Run Networks builds mission-driven intelligent systems for national security and other high-stakes government missions. We apply commercial innovation to operational missions, an approach we call reverse dual use. As mission timelines outpace the acquisition system, we work at startup velocity with enterprise credibility. We build solutions that a program office can stand behind, and that operators want to use.

This role supports a federal civilian agency modernization program that applies AI and modern data tools to 24/7 operations, readiness, and crisis response.

About the job
This program is changing how operational information is collected, triaged, and escalated across teams that work around the clock. The program opens with a discovery phase, so before anything is built, you help decide what to build alongside the people who will use it.

You build the machine learning and AI systems behind that work. You own each model from the data through deployment and monitoring, and you work directly with the analysts and operators who will rely on it. You are expected to explain in plain language what a model does and where it can be wrong.

What you'll do

 * Build models people use: Design, train, and retrain machine learning and generative AI capabilities, including retrieval-augmented generation and agent-based workflows, to solve real operational problems.
 * Take prototypes to production: Move models into secure production environments using Git, CI/CD, and containers, and support the security authorization the agency requires before anything goes live.
 * Measure and monitor: Build evaluation, drift detection, and explainability checks that catch a bad output before an end user does.
 * Integrate with the program's systems: Connect models to applications, dashboards, and repositories through APIs and services.
 * Work with the people who will use it: Sit with analysts and operators to define the question, the data, and what a model is allowed to decide versus what needs human review.
 * Document the system: Record assumptions, limitations, evaluation results, and operating procedures clearly enough that someone other than you could run the system.