Gen AI Engineers

Accellor•San Francisco, CA

About The Position

Accellor is seeking a GenAI Engineer to advance how software is designed, built, tested, and delivered using frontier AI models and coding agents. The engineer will work with models such as OpenAI and Anthropic Claude, as well as AI development environments such as Cursor, combining software engineering fundamentals with prompt engineering and agent orchestration to improve development productivity, software quality, and cost efficiency. The role suits engineers who enjoy building working software, evaluating new approaches, and turning successful experiments into reliable engineering practices. Recent graduates with strong projects and demonstrable skills are encouraged to apply. The engineer will develop and deliver software with AI-assisted tools, design and evaluate prompts and reusable workflows, orchestrate coding agents, and measure improvements in engineering efficiency. They will take ownership of the quality and delivery readiness of code, including AI-generated changes, and collaborate with product, engineering, and quality teams. Success is demonstrated through sustained contributions to software delivery, engineering quality, and team effectiveness, with performance expectations reflecting the role’s scope, complexity, and ownership.

Requirements

0–4 years of relevant experience in software engineering, AI engineering, or applied AI development; academic, personal, and open-source projects are also considered.
Proficiency in at least one programming language, such as Python, TypeScript/JavaScript, Java, Go, or C#.
Hands-on experience applying frontier models, including OpenAI models or Anthropic Claude, to software development.
Practical experience with Cursor or comparable AI coding environments.
Strong prompt engineering skills, including task decomposition, context preparation, structured outputs, and iterative evaluation.
Working knowledge of Git, pull requests, debugging, automated testing, APIs, and CI/CD.
Ability to understand existing codebases and independently assess the correctness of AI-generated code.
Clear communication, collaborative working habits, and ownership of engineering outcomes.

Nice To Haves

Experience developing workflows with multiple coding agents, including delegation, handover, and integration.
Familiarity with model APIs, tool calling, Model Context Protocol (MCP), and agent orchestration frameworks.
Experience evaluating prompts and models using defined datasets, acceptance criteria, and automated checks.
Experience monitoring AI usage, engineering productivity, and cost efficiency.
Exposure to cloud platforms, containers, deployment automation, and application observability.
A portfolio of working software demonstrating effective use of AI tools, supported by tests and technical documentation.

Responsibilities

Develop features, resolve defects, refactor applications, and create automated tests using AI-assisted development tools.
Use Cursor and other coding environments for repository exploration, implementation, debugging, testing, and code review.
Translate product requirements into engineering tasks, technical approaches, and acceptance criteria.
Own code quality and delivery readiness, including changes produced by AI models or agents.
Collaborate with product, engineering, and quality teams to deliver reliable software releases.
Design and maintain prompts, repository instructions, context strategies, and reusable workflows for software engineering tasks.
Provide models with relevant architecture, dependencies, coding standards, constraints, and expected outcomes.
Establish repeatable evaluations to assess correctness, consistency, requirement coverage, and the need for human intervention.
Improve prompts and workflows based on test results, code reviews, and delivery outcomes.
Build coordinated groups of coding agents to support planning, implementation, testing, review, and documentation.
Design task delegation, handover, takeover, and recovery workflows with clear ownership and shared task state.
Manage parallel execution, task dependencies, isolated workspaces, and integration conflicts.
Define appropriate tool access, execution boundaries, and human review checkpoints.
Ensure agent-produced changes can be traced, reviewed, tested, and integrated into the codebase.
Select models and tools based on task requirements, quality, latency, and cost.
Optimize context size, token consumption, model routing, retries, and agent execution overhead.
Establish baselines and measure changes in development throughput, delivery time, review effort, and AI expenditure.
Report weekly and monthly delivery trends, including pull requests raised and merged.
Share validated practices and reusable tools that improve the wider engineering team’s effectiveness.
Deliver meaningful, well-scoped changes with predictable execution and clear ownership; contribute to task completion and improve time to review-ready code and pull request cycle time.