AI/MLSpecialist

This is an illustrative working day, not an official DCWF schedule. Exact duties depend on the application and employer.

DCWF describes work rather than setting salaries. There is no single US salary figure for every position mapped to role 623.

// A related US occupation

BLS reports this annual median for May 2025 and projects 35% employment growth from 2025 to 2035 for data scientists. This is context for a related occupation; it is neither an AI engineer salary estimate nor a DCWF 623 forecast.

Read the BLS data and education guidance

// Compare real opportunities

• Compare advertised base salary, location, experience level and the work described. Separate internships and junior roles from senior or research positions.
• AI engineering vacancies may emphasise software development, data science or research. Their pay and entry requirements can differ.
• For federal jobs, read the announcement’s grade, salary range, eligibility, education and experience requirements.

The route below is suggested career guidance, not an official DoD qualification pathway. Build evidence that you can program, work with data and test whether a solution is useful and safe.

// Grades 9–12: start where you are

• High school: Take available computer science and math courses. Build algebra first, then probability, statistics, calculus and linear algebra where offered. Learn Python, spreadsheets and simple data visualization.
• Clubs and projects: Try coding, robotics or science projects. Explain the question, the data, the result and what you would improve. You do not need an expensive computer to start with small datasets.

// Considering or studying a degree

• Compare computer science, software engineering, data science, mathematics, statistics and related engineering degrees by their actual courses, rather than the word “AI” in the title.
• Look for programming, algorithms, databases, linear algebra, probability, statistics, machine learning, software testing and security. Add ethics and privacy study.
• If you are already studying cybersecurity or IT, consider programming, math and ML electives and an AI project. Check prerequisites with your university rather than assuming your degree alone covers the role.
• Use team projects, supervised research, placements and internships to practice building and evaluating applications. A master’s or doctorate may be relevant for research-focused vacancies; check each employer’s requirements.

// Early career or skills-first route

• If you work in software, add data preparation, ML evaluation and model integration. If you work in data analysis, strengthen programming, testing and application development.
• If you work in IT or security, use synthetic logs to explore anomaly detection. Learn why unusual activity is not automatically malicious and why human review matters.
• Search for software or data internships, junior developer and junior data roles alongside AI/ML positions. Treat these as ways to build relevant experience, not guaranteed promotions into AI engineering.
• Publish a clear portfolio with code, instructions for running it, where the data came from, results, limitations and your own contribution. Match applications to duties you can demonstrate.

// Portfolio project: a small phishing classifier

• Use a public dataset with suitable reuse permission or teacher-created fictional messages. Do not collect classmates’ private emails or personal information.
• Start with a simple rule-based starting model, then compare it with a basic text classification model. Keep training and test examples separate.
• Measure both false alarms and missed phishing examples; explain why each matters. Inspect errors instead of reporting accuracy alone.
• Build a small interface using example inputs. Record the data source, code version, test method, security considerations and limitations. Label it a learning prototype, not a production email security product.

Certifications can support learning or demonstrate knowledge, but a course certificate does not establish that you can build reliable AI software.

// DoD proficiency is based on the work

The coding guide uses Basic, Intermediate and Advanced proficiency levels. Basic work involves routine situations with frequent guidance; Intermediate work includes more complex situations with periodic guidance; Advanced work requires deeper expertise, little guidance and the ability to guide others. These describe the work expected in a position, rather than a sequence of school qualifications.

// Before paying

• Check whether target employers request the credential, and whether it assesses the skills you need.
• Distinguish a completion certificate from a certification that includes an exam or other skills assessment.
• For a DoD position, confirm the assigned work role, proficiency level and applicable qualification options with the employer.

Check official DoD 8140 guidance

Try these US search terms, then compare the actual responsibilities with AI development work. These links open search results, not selected or endorsed vacancies.

// Read the advert carefully

• Look for duties involving AI application development, model building or adaptation, testing and integration. Some “AI” titles describe sales, product management or simply using AI tools.
• Check programming and math prerequisites, experience expectations, internship eligibility and whether the role is genuinely junior.
• Federal and defence vacancies may specify citizenship, clearance or other eligibility conditions. Apply the requirements stated in each advert.

// You’ll probably enjoy this if you…

• Like writing code and finding out why it fails.
• Enjoy math, patterns and experiments, including results that challenge your first idea.
• Want to build tools people can use and explain when they should not rely on them.
• Care about security, privacy and the consequences of incorrect predictions.

// Think carefully if you…

• Only enjoy using chatbots and do not want to learn coding or work with data.
• Dislike repetitive debugging, data cleaning or comparing experiments.
• Expect every problem to have one certain answer. AI systems can fail in ways that require careful investigation.

This related Data/AI role checks AI systems through testing and validation, examines their resilience, security and trustworthiness, and reports concerns to leaders.

For an AI developer, this is the connection between building a solution and gathering evidence that it works safely. It is a distinct work role, not an automatic next career step.

The role code, name and Data/AI classification are confirmed in the coding guide’s Appendix B, Tables B-1 and B-2 (printed pages 19–20).

Open the official dashboard and select role 672

This related Data/AI role helps people understand AI risks and assesses technical and societal consequences throughout a solution’s lifecycle.

For an AI developer, this connects design choices with responsible use, risk assessment and the effects on people. It is a distinct work role, not a certification pathway.

The role code, name and Data/AI classification are confirmed in the coding guide’s Appendix B, Tables B-1 and B-2 (printed pages 19–20).

Open the official dashboard and select role 733