Technical leader of the AI solutions practice: accountable for the architecture and quality of solutions the team takes from pilot to production, and for the technical management of the team.
Scope of responsibility
Technical architecture of AI solutions: choosing the approach (RAG, fine-tuning, agentic pipelines, classical ML) and assessing the cost / quality / time-to-deploy trade-offs
Decomposing business requirements into a technical backlog; effort estimation and building estimates with appropriate buffers
Technical management of a team of 5-10 people: task allocation, code review, mentoring, competency development
Ownership of delivering projects to a measurable business outcome rather than to a demo
Establishing the quality loop for models: metrics, evaluation sets, regression testing of prompts and pipelines
Technical communication with clients: defending architectural decisions, participating in presales and proposal development
Must-have requirements
5+ years in ML, including 2+ years in a technical leadership role (team lead, tech lead)
Hands-on experience taking LLM solutions to production: RAG, vector stores, agent orchestration, working within context and model constraints
Production-level Python; solid understanding of API integrations, queues, and asynchronous pipelines
Experience with cloud platforms (Azure / AWS / GCP), containerisation, and basic CI/CD
Estimation and planning skills: able to produce a realistic estimate and explain what it consists of
Ability to hold a technical conversation with business stakeholders without falling into jargon
English at Upper-Intermediate level or above (documentation and client-facing work)
Nice to have
Experience in a consulting or project-based delivery model (multiple parallel clients, fixed scopes)
Experience building LLM output quality assessment processes (LLM-as-judge, human-in-the-loop)
Experience with as-is / to-be process mapping and identifying automation opportunities
Experience introducing AI-assisted development practices (agentic coding tools, conventions) to a team
Experience with self-hosted model deployment and its cost economics
Soft skills and management profile
Takes ownership of the team's results, not only of their own code
Able to tell a client "this won't work" and propose an alternative
Develops people: walks through the reasoning behind decisions rather than simply correcting them
Comfortable working with unstable requirements — the normal condition of AI projects
Willing to run difficult performance conversations, not only supportive mentoring
