Staff AI Deployment Engineer, Altana

• 8+ years in customer-facing technical delivery — implementation, engineering, professional services, technical program management, or product management with hands-on scope
• A demonstrated track record of shipping complex technical projects with external stakeholders, on a real deadline, with real consequences
• Strong communicator and coordinator: you can run a customer working session in the morning and drive internal alignment across three teams in the afternoon
• Product instincts — you naturally ask what problem this actually solves and what the smallest thing that solves it is
• Comfortable with ambiguity and partial information; you make progress before the picture is complete
• Bias toward automation and repeatability over heroics; you'd rather fix the process than run it manually again
• Working proficiency in Python or a comparable language: you can read a codebase, write and debug scripts, build integrations, and ship small features — you don't need to be a systems engineer
• Fluency with SQL and comfort exploring and validating data to answer your own questions
• Practical experience with AI-assisted coding tools (Claude Code, Cursor, Copilot, or similar) as a core part of how you work — you use them to move faster on unfamiliar codebases, prototype integrations, and automate the repetitive parts of delivery, and you know how to review what they produce
• Hands-on experience building or deploying agentic AI systems — LLM-powered agents, tool and function calling, retrieval-augmented generation, or multi-agent orchestration
• Skill in prompt engineering and agent evaluation: you can design prompts and evals, diagnose why an agent fails, and iterate toward reliable production behavior
• Familiarity with agent frameworks and protocols (e.g., MCP, LangChain, or similar) and with wiring agents to external tools, APIs, and data sources
• Familiarity with APIs (REST, Swagger/OpenAPI), Git-based workflows, and cloud-hosted systems
• Enough understanding of data pipelines, data modeling, and ML systems to scope work, diagnose issues, and hold a credible conversation with the engineers who build them

Desirable

• The Implementation team is the tip of the spear in getting Altana's platform into production inside our customers' businesses. We own the path from signed contract to realized value: standing up integrations, wiring Altana into the customer's existing business processes, and making sure the insight we deliver actually changes decisions on the customer's side
• We are a delivery organization with a technical core. We work directly with customer stakeholders — from trade compliance analysts to enterprise IT — and we work just as closely with Altana's engineering, data science, and product teams to translate what we learn in the field into what gets built next
• Design and deploy agentic AI systems
• Design, build, and deploy agentic AI workflows — multi-step agents, tool and function calling, and automations — that power the customer's business processes and act on real customer data
• Ground agents in customer context: connect them to the right data sources, APIs, and systems of record, and define the tools, permissions, and guardrails they operate within
• Write and refine the prompts, system instructions, and evaluation criteria that make an agent reliable enough to trust in production
• Build the orchestration and human-in-the-loop checkpoints that keep agentic workflows safe, auditable, and correctable in a customer's environment
• Monitor, evaluate, and iterate on deployed processes — tracking accuracy, cost, and latency, and tuning them as the customer's data and needs evolve
• Own technology implementation and delivery
• Lead customer implementations end to end: scoping, configuration, integration, validation, go-live, and handoff to steady-state support
• Configure and deploy Altana's platform capabilities against real customer data and real customer workflows, resolving the gap between what the product does today and what this customer needs
• Build the connective tissue — API integrations, data handoffs, automations, and internal tooling — that makes an implementation work, using AI-assisted development to move quickly
• Drive implementations to a definition of done that includes adoption, not just deployment: the customer's team is using it, in their process, without us in the room
• Engage customers as a trusted technical partner
• Serve as the primary technical point of contact through implementation — running working sessions, demos, or technical deep dives with stakeholders from analyst to executive
• Elicit requirements from customers who often can't fully articulate them yet, and translate business problems into concrete technical scope
• Communicate credibly in both directions: explain platform behavior, data quality realities, and trade-offs to non-technical audiences, and represent customer constraints faithfully back to Altana
• Manage expectations, timelines, and scope with the directness that keeps large enterprise programs on track
• Coordinate across internal teams
• Orchestrate the internal effort behind each implementation, pulling in engineering, data science, product, and customer success at the right moments and keeping them aligned on a single plan
• Escalate and unblock: identify what only another team can solve, make the ask precisely, and follow it to resolution
• Feed field learning back into the product organization — patterns across customers, recurring friction, gaps worth building
• Build reusable playbooks, templates, and documentation so the next implementation is faster than the last
• Apply product management judgment
• Own the implementation roadmap for your accounts: sequence work by customer value and delivery risk, and make explicit calls about what ships now versus later
• Write clear requirements, specs, and acceptance criteria that engineering and data science can build against
• Prioritize against customer value delivery timelines and cooperate with the product team on milestones and deadlines
• Define what success looks like for each implementation in measurable terms, and track against it