Principal AI Engineer

• Deep software engineering and architecture experience (system design, testing, refactoring, code review) at a level where you confidently judge and correct anything an AI produces.
• Demonstrable AI-first delivery experience: you already ship real, production work primarily through agentic tools. Be ready to show how.
• Fluency with modern AI coding environments (Claude Code, Cursor, or equivalent), including plan and agent modes.
• Specification-driven, context-engineering discipline: you frame the problem and the design before you let an agent run.
• Sound judgment on when not to trust the agent: you catch hallucinated APIs, silent logic errors, and LLM antipatterns in generated code.
• Autonomous — take a vague goal, scope it, and drive it to delivery without waiting to be told the next step.
• Business-oriented and business-driven — optimize for client and business outcomes, not technical elegance for its own sake.
• Ownership — accountable for outcomes end-to-end, including what an agent produced under your direction.
• Decisiveness under ambiguity — sound calls with incomplete information, adjusted as reality lands.
• Strong communication — explain technical decisions to clients and non-engineers; make the implicit explicit for teammates and agents.
• Pragmatism over dogma — pick the approach that ships quality fastest, not the most fashionable one.
• Mentorship — raise the people around you, not just your own output.

• Hands-on experience with spec-driven frameworks (BMAD, OpenSpec, Superpowers).
• Brownfield AI transformation: bringing agentic delivery into a legacy codebase.
• Building graph and MCP context layers so agents can pull real project context.
• Experience designing and shipping AI-powered product capabilities (LLM, RAG, agents).
• Token-cost and FinOps thinking for agentic engineering.

Principal-level, hands-on engineering role: you deliver through an AI-native SDLC — define the work and quality bar (architecture, specifications, acceptance criteria), use AI agents for the heavy lifting, and remain accountable for what ships. This is an Engineer role, not advisory: you write and ship the work yourself, amplified by agents. Stack-agnostic.

• Own end-to-end delivery of complex features and services through an agentic, specification-driven workflow, with architecture and acceptance criteria defined up front and AI executing the build under your direction.
• Architect systems that are legible to both humans and agents, and make the design and tooling calls that let a team deliver AI-first.
• Orchestrate AI coding agents (Claude Code, Cursor agent layer, Kiro and similar) across targeted edits and large orchestrated refactors and builds; selecting the right tool for the job is part of the craft.
• Practice agentic TDD: tests and specifications drive the agent, and quality is engineered in rather than bolted on.
• Run AI-first code review on agent-generated changes before human review, so reviewers are not overwhelmed by volume.
• Treat the token and cost economics of agentic delivery as an engineering constraint you actively manage.
• Shape AltexSoft's AI-native SDLC practices and mentor teams moving from manual to AI-first delivery.
• Keep humans in the loop where judgment matters: architecture, security, ambiguous requirements, and the final quality call.