AI Trainer: Code Generation

About the role

This position is centered on the analysis and refinement of multi step reasoning trajectories derived from real production code to enhance the reliability of large language model code generation. It represents a sustained engineering initiative aimed at improving cross platform execution quality rather than a short term labeling effort. The hire will work closely with production repositories to identify weak steps in logical sequences and convert them into robust, production grade outputs. Success in this role requires a deep understanding of how code behaves across multiple environments and platforms. The individual will shape trajectory structure and clarity by evaluating reasoning quality under varied constraints. Strong written communication will be used to document findings and clearly articulate proposed refinements to engineering teams. This role operates as part of a focused group dedicated to building long term evaluation capabilities for code generation systems.

Key facts

What you'll do

You assess model generated reasoning sequences for logical consistency and surface weak steps in multi step problem solving using systematic examination of execution paths. Trajectories sourced from real production repositories are refined to create production grade outputs that meet rigorous engineering standards and pass practical validation. Evaluation of reasoning quality across varied programming environments guides improvements to trajectory structure, ensuring clarity and alignment with target outcomes. Complex code paths are examined to confirm that reasoning aligns with practical implementation patterns, constraints, and deployment considerations. You collaborate with engineering teams to translate identified issues into concrete improvements in how models plan and execute code. Instrumentation and tooling are used to trace reasoning steps, isolate failure points, and measure the impact of refinements over time. Insights from these evaluations feed back into trajectory design, prompting iterative adjustments that reduce brittle behavior and increase robustness. You contribute to the definition of evaluation criteria that reflect real world engineering expectations and measurable quality metrics. This role emphasizes engineering judgment over surface level labeling, requiring careful navigation of ambiguous or under specified problems. The work directly influences the reliability of code generation systems that operate across multiple languages and platforms.

Requirements

You must be proficient in at least two mainstream programming languages such as Python, C++, Java, TypeScript, or JavaScript to analyze and discuss generated code with confidence. You must demonstrate real world development experience in backend systems, frontend applications, algorithms, testing, or infrastructure to understand how code integrates into larger systems. Comfort with reading and reasoning through large GitHub repositories is required to navigate complex codebases effectively and locate relevant implementation patterns. Strong written communication skills are necessary to document findings and proposed refinements clearly, enabling actionable feedback for engineering teams. A degree is required as part of the eligibility criteria for this position, ensuring a baseline level of formal training. You should be able to break down unfamiliar problems into logical steps and explain your reasoning in plain language during interviews and code reviews. Previous exposure to code generation tools, testing frameworks, or static analysis is valued but not mandatory. The ability to work independently and own the full lifecycle of a trajectory evaluation task is essential for success in this role.

Practical notes

Typical interview steps

Hiring for engineering roles usually starts with a recruiter screen, followed by one or two technical rounds. Candidates often solve a coding problem, discuss past projects, and answer system design questions. Some loops include a take-home task. Final rounds typically cover team fit and give candidates a chance to ask questions. Interviewers look for how you break down an unfamiliar problem, not just whether you reach the answer. Practicing a few problems aloud and reviewing your own past projects are the best preparation.

Good to know

The role emphasizes engineering judgment over surface level labeling. Contributors work directly with complex code paths across platforms. Tools commonly used include code hosting services and language ecosystems. This role operates as part of a focused engineering group building long term evaluation capabilities.

Questions to ask

Useful questions for the interview: what a typical week looks like, how work is assigned, what tools the team uses, and how feedback works. Asking how the role has changed recently and what the team wishes it had known when joining is also reasonable. Questions about the manager's priorities are especially valued.

Career growth

Engineering careers usually progress from individual contributor to senior, staff, and principal levels. Some engineers move into management and lead teams of five to twenty people. Others stay on the technical track. Growth follows demonstrated impact, not tenure alone. A typical engineering ladder has clear levels with defined expectations for scope, quality, and mentorship. Moving up usually requires owning outcomes end to end rather than completing assigned tickets.

About the company

ability and creative control as you do about technical novelty. - 3+ years of engineering experience

Bonus: experience with creative tools, generative models, or node-based interfaces. WORKING AT FLORA OFFICE POLICY We work in-person at the Domino Refinery building in Williamsburg, Brooklyn - with views of the East River.