Agentic AI Technical Trainer

Technical AI trainer (contractor, remote and on-site)Teaching engineers to build agentic systems is not the same as demoing them. It takes someone who has built and shipped the thing, and who can then take a room of engineering, IT and research staff from first principles to working systems, live, in the client's environment, on their keys. Rayify delivers technical AI training to financial institutions, and we are looking for contract trainers who can run that room.About RayifyRayify designs agentic AI workflows and decision intelligence systems for regulated financial services, covering insurance, credit and private markets. Our systems deploy inside clients' own infrastructure and train on their proprietary data. We build them to be auditable for boards and regulators, so outputs stay traceable and secure and clients can manage risk while still getting insight from their data.To keep in touch, follow Rayify on LinkedIn: https://www.linkedin.com/company/rayifyWhat you'll doDeliver our technical workshop track to engineering, IT and research teams at financial institutions: hands-on sessions running from two hours to full build days, virtually and in personRun live builds and demos in front of technical audiences, covering agents, orchestration, tool use, retrieval and guardrailsAdapt each session to the client's stack, sector and workflows, including custom workflow automation engagements where you map priority operational challenges and build against their infrastructureFeed what you learn in the room back into the curriculumThe workshops you'll deliverWe teach across the major agent ecosystems: Anthropic (Claude Code, the API and the Agent SDK), Microsoft, OpenAI, and open-weights models for clients who need everything to run inside their own environment. The choice follows the client's existing estate, and the principles transfer between all of them. You do not need to cover every ecosystem or every workshop. Tell us where you are strongest.Coding agents for engineering teams: turning a general-purpose coding agent such as Claude Code, GitHub Copilot or Codex into a team's tool, with scoped sub-agents, skills and shared memorySub-agent orchestration and observability: custom sub-agents, orchestration patterns, and observability with telemetry for teams working interactively in a coding agentCode-driven agent orchestration: reliable, stateless multi-agent architectures built on the major agent SDKs, including linear sequences, concurrent execution and smart routingRAG and context engineering: building RAG pipelines, optimising retrieval quality beyond naive search, and context engineering techniquesSelf-hosted and open-weights deployment: choosing and serving open-weights models inside the client's own infrastructure, and what changes about orchestration, tool use and evaluation when you doAgentic governance, guardrails and security: the agentic threat landscape, enforceable security controls, and red-teaming methodologies for autonomous systemsEvaluating agentic systems (coming soon): designing eval sets for agents rather than chatbots, covering task and grader design, rubrics and LLM-as-judge, trajectory and tool-call evaluation, regression testing, and turning eval results into a go/no-go signal that model risk and security review will acceptCustom workflow automation and engineering hackathons: half-day to full-day builds against client infrastructure, deployed live where access allows, with closing demos and peer reviewWho this suitsMust haveYou have shipped LLM or agentic systems, not just prototyped them ideally within financial services institutions: agent orchestration, tool use and MCP, retrieval, evaluations, guardrailsDeep hands-on fluency with the Claude ecosystem, meaning Claude Code, the API and the Agent SDK, plus strong Python or TypeScriptA track record of delivering hands-on technical training to engineering audiences, in workshops and live builds rather than talksComfort building live in an unfamiliar client environment, and recovering gracefully when the demo gods misbehaveExceptional communication skills. You can pitch the same material to a staff engineer and to a technically curious operations lead, and translate between themNice to haveFinancial services or other regulated-industry experience: security review, data boundaries, audit trailsExperience designing evaluations for agentic systemsDepth in agentic security, including threat modelling, red-teaming and enforceable controlsExperience developing training curricula as well as delivering themEngagement termsContractor arrangement, booked per session or per dayRemote and on-site deliveryWe invite applicants from around the worldQuote your hourly or day rate when you applyWhy joinYou get regular paid delivery into financial institutions without having to build a curriculum from scratch. You also join a network of domain experts and engineers, where training engagements often lead into deployment work.How to applyApply with your CV, GitHub or LinkedIn, a note on the audiences you have trained and the most substantial agentic system you have shipped, and your rate.To keep in touch, follow Rayify on LinkedIn: https://www.linkedin.com/company/rayifyWe consider all qualified applicants without regard to legally protected characteristics, and we provide reasonable accommodations on request.