Snr Code Reviewer - C

About OpenTrain

OpenTrain aggregates data-labeling and AI-training jobs from many AI companies and labeling platforms into one job board so you can find this work without hunting across dozens of sites. Creating an OpenTrain account is free, and applying takes only a few minutes.

We connect skilled reviewers and annotators with short- and long-term projects that help train the next generation of AI systems.

About AI training work

Human reviewers and annotators prepare and check examples that AI models learn from — everything from labeling images to rating generated code. That human validation is how models become accurate, safe, and useful.

This role focuses on code-level annotation: verifying generated C code, correcting reviewer ratings, and ensuring each labeled example meets quality and safety expectations.

The Role

You will audit annotator reviews of AI-generated C code (subject matter: AI-generated C code response annotation) and ensure every review aligns with our quality rubric. Work is labeled as COMPUTER_PROGRAMMING_CODING and uses other labeling software.

For each submission you will compile and run the snippet in a sandboxed container, verify it meets the prompt, test under multiple build flags, check performance and security, correct any mis-ratings, and supply concise feedback to improve reviewer accuracy.

What you'll do

• Compile and run submitted C snippets in a sandboxed/containerized environment and verify prompt compliance and correct behavior.
• Test code under multiple build flags and toolchains (GCC/Clang), and validate across relevant compiler extensions.
• Use sanitizers (ASan/UBSan/TSan), Valgrind, debuggers (gdb/lldb), and unit-test frameworks to detect defects and confirm fixes.
• Identify undefined behavior, buffer overruns, memory leaks, double-frees, dangling pointers, data races, and other safety issues.
• Assess performance and profiling data (perf, VTune) for cache-friendliness and low-level tuning opportunities where applicable.
• Correct mis-ratings, supply concise, rubric-aligned feedback, and ensure every review follows structured QA checklists.
• Work with CI/CD flows and code-review platforms to validate fixes and reproduce issues in Dockerised toolchains.

Requirements

• 7+ years of professional C development, QA, or dedicated code-review experience.
• Deep knowledge of modern C (C11/C18), the C standard library, and common compiler extensions (GCC/Clang).
• Experience with build systems such as CMake, Meson, and Make.
• Proven memory- and concurrency-analysis skills: detect leaks, double-frees, dangling pointers, UB, data races; strong grasp of atomics and locks.
• Advanced use of sanitizers (ASan/UBSan/TSan), Valgrind, gdb/lldb, and unit-testing frameworks (Unity, CMock, Catch2, etc.).
• Performance tuning and profiling experience (perf, VTune) and low-level optimization across architectures.
• Secure-coding expertise: identify/migrate CWE/OWASP issues (buffer overflows, integer wrap, format strings) and understand exploit mitigations (stack canaries, ASLR).
• Comfortable with Dockerised toolchains, CI/CD pipelines, and code-review platforms (GitHub/GitLab PRs).
• Proof-of-work validation experience compiling and running code in sandbox/container environments to verify functionality and prompt compliance.
• Familiarity with rubric-based scoring, checklist-driven reviews, and ticketing/annotation tools (Jira, Asana).
• Excellent written English (B2+ CEFR) for clear, concise feedback and mentoring.

Who should apply

This role is aimed at senior C developers, security-minded systems engineers, and experienced QA/code-review professionals who enjoy hands-on debugging and clear written feedback.

Nice-to-have: background in LLM evaluation, RLHF pipelines, or prior AI/ML data-labeling projects — helpful but not required.

How it works, hours & pay

This is a remote, part-time contractor role with a minimum commitment of 20+ hours per week. The position is worldwide/remote and uses PAY_PER_HOUR compensation at USD 24 per hour.

OpenTrain lists and aggregates the role; applying through OpenTrain is free and quick. The project will use other labeling software and provide access to sandboxed environments and tooling required to perform reviews.