AI/ML engineer (Part Time) - WFH
About the Company
OnSite IT is an IT support and consulting company based in Ocala, Florida. The company supports software development teams and clients worldwide through a small, remote team.
About the Role
OnSite IT is looking for an experienced AI/ML Engineer to provide technical guidance on its AI products and LLM systems for fewer than five hours per week. This part-time role is suited to an engineer who already has a full-time position and wants a limited, ongoing commitment focused on architecture, technical reviews, and production readiness. You will work directly with the founders and engineering team to evaluate the product roadmap, guide AI architecture decisions, review code, and identify potential problems before they affect production. Most collaboration takes place through written reviews, architecture notes, and chat, with no on-call responsibilities or required attendance at daily standups.
Key Responsibilities
Review the AI and product roadmap with the founders every few weeks and recommend which initiatives to prioritize. Design architecture for AI features and LLM systems, including model integration, data retrieval, prompt testing, and failure handling. Evaluate system reliability, response times, model output quality, and API costs to identify potential production issues. Review code and pull requests, providing actionable feedback on security, scalability, maintainability, and overall system design. Recommend appropriate model providers, vector databases, frameworks, deployment approaches, and production monitoring tools. Identify unnecessary tools or infrastructure and help the team make practical technology choices within its budget. Participate in project planning to assess technical scope, identify potential delays, and highlight areas that may require additional effort. Answer engineers' technical questions through clear written guidance. Write code when doing so is more efficient than explaining the required changes. Guide and mentor engineers through technical feedback, architectural recommendations, and development best practices.
Required Qualifications
Experience building and operating LLM systems used by real users in production. Practical experience troubleshooting production LLM issues, including poor output quality, slow responses, increasing API costs, and prompts that become less effective when models change. Strong full-stack engineering background with experience shipping and maintaining production systems. Experience designing software architectures and making technical tradeoffs for small teams with limited budgets. Familiarity with AWS, Google Cloud Platform (GCP), or Microsoft Azure. Knowledge of CI/CD pipelines and standard software deployment practices. Strong written communication skills, with the ability to provide clear, actionable technical guidance. Previous experience guiding or mentoring other engineers.
Preferred Qualifications
A degree in Computer Science, Engineering, or a related field is welcome but not required. Practical experience and demonstrated production work are the primary considerations. Location & Work Arrangement Work arrangement: Remote. Employment type: Ongoing part-time engagement. Time commitment: Fewer than 5 hours per week. Schedule: Flexible, with work completed asynchronously and on your own schedule. On-call duties and daily standup attendance are not required. Additional Terms This role is intended for professionals who wish to retain their current full-time employment while taking on a small additional commitment. Candidates should confirm that their current employment agreement permits outside work and disclose any potential conflicts of interest before starting. Employer confidentiality must be respected throughout the engagement.
How to Apply
Please describe one LLM system you have taken into production. Explain what the system did, what initially went wrong or performed poorly, and how you addressed the issue. A short paragraph is sufficient. Include your available weekly hours and expected rate when applying.