Mohamed Z.-RL Environments & Benchmark Engineer
Check rate
Experience
RL Environment & Benchmark Author
AfterQuery
- Authored SWE-bench-style evaluation tasks across a large portfolio of production codebases, each shipping a Docker environment, test harness, and verified solution script; graded against reward-hack resistance, behavioral verification, and mergeability rubrics.
- Designed and evaluated long-horizon frontier-SWE agent tasks - multi-step objectives spanning several files and stages, where the agent must hold context across a long trajectory and where partial progress must be distinguishable from a real solve by the verifier.
- Built PyTorch / GPU-kernel systems tasks in a containerized harness format (CUDA base images, verifier timeouts, resource-bounded environments), each validated locally against both a no-op agent (reward 0, all target tests fail) and the oracle solution (reward 1, all tests pass) before submission.
- Calibrated task difficulty to a target solve band by coupling multiple silent defects across files, adding decoy competence paths and misleading symptoms - writing symptom-only but contract-complete instructions that withhold the cause without withholding the spec.
- Constructed C/C++ and Rust security benchmark repositories with planted memory-safety bugs, calibrated against CyberGym/ARVO/OSS-Fuzz corpora so that frontier models resolve only a small fraction of planted defects; diagnosed submitted repos for cross-solve vulnerabilities and isolation failures.
- Designed agent-evaluation prompts across multiple task domains, trajectory modifiers, and safety axes for computer-use agent scoring.
- Audited peer-authored tasks for verifier over-specification and false-negative gates; built and maintained a team task-tracking dashboard.
- Sustained a high weekly authoring throughput against hard deadlines while at university, across multiple concurrent projects.
Freelance ML Systems Engineer
Afterquery
- Optimized a 7B-parameter LLaMA fine-tuning pipeline under VRAM constraints (4x A100 40GB).
- Wrote a custom CUDA extension via pybind11/ATen fusing LayerNorm and RoPE, cutting kernel launch overhead ~30%.
- Implemented a custom autograd.Function for memory-efficient cross-entropy that recomputes gradients in the backward pass instead of materializing the full logit matrix.
Net effect: 18% lower peak memory, 2x batch size without OOM on long runs.
Full-Stack & Data Contributor
Various
- Next.js + PostgreSQL feature work: API routes, backend services, SQL migrations, Redis caching and cache-invalidation strategy on high-traffic endpoints.
- Structured evaluation and labelling for LLM and vision pipelines (Alignerr / Outlier / DataAnnotation).
Software Engineering Intern
Mattel S.A
- Internal tooling across the SDLC; code review and cross-project troubleshooting.
Industry experience
See where this freelancer has spent most of their professional time.
Experienced in Information Technology and Manufacturing.
Business area experience
See which departments and functions this freelancer has contributed to most.
Experienced in Information Technology and Product Development.
Summary
Benchmark and RL-environment engineer. I build the artifact these teams sell: containerized, verifier-backed tasks with golden solutions, calibrated so frontier models fail in a target band and can't reward-hack their way to a pass. ~1 year authoring and auditing environments across long-horizon frontier-SWE, PyTorch/GPU-kernel, and memory-safety security domains, plus agent-trajectory grading. Backed by low-level C/C++/CUDA work and a competition mathematics background (double PAMO bronze medalist).
- Pan-African Mathematics Olympiad (PAMO 2022, 2023) - Double Bronze Medalist
- Presidential Prize for Sciences (Rally 2022, 2023) - Bronze Medal
- Scholarships: Türkiye Bursları (Turkish Government Scholarship); BMCI Bank Scholarship
Skills
- Rl Environments & Evals: Long-Horizon Agent Task Design; Task Authoring And Difficulty Calibration; Docker Environment Design; Verifier And Test-Harness Engineering (Fail To Pass / Pass To Pass); Golden Solution Scripts; Reward-Hack Resistance; Behavioral Verification; Solve-Rate Banding; Agent Trajectory Grading And Safety-Axis Rubrics; Task Auditing (Verifier Over-Specification, Cross-Solve Leakage, Clone Clusters, Grep Isolation).
- Ml Systems: Pytorch Internals, Custom Autograd.Function, Cuda/Tensor Extensions, Aten-Level Ops, Pybind11, Kernel-Level Failure Modes (Views/Aliasing/Strides/Dtype/Accumulation Order), Llama Fine-Tuning Scripts.
- Data Science & Analytics: Python, Machine Learning, Statistics, Data Analysis, Experimentation, Pandas, Data Modeling, Reporting, Sql, Snowflake.
- Qa & Testing: Quality Assurance, Software Testing, Automation Testing, Test-Case Design, Regression Testing, Bug Tracking.
- Gpu Programming & Graphics: Cuda, Webgpu, Glsl, C++.
- Mechanical Design Engineering: 5+ Years Of Hands-On Experience Focused On 3d Cad Modeling And Fea; Multi-Platform Proficiency In Solidworks, Nx, Creo, And Catia; Component Design; Requirements Definition; Technical Drawing Creation And Interpretation; Specifications And Tolerances; Stress Analysis And Fea Principles; Engineering Validation; End-To-End Workflow From Requirements Through Modeling, Analysis, Validation, And Documentation.
- Engineering Documentation & Communication: Comprehensive Technical Drawings And Engineering Documentation; Precise Requirements/Specification Writing; Meticulous Attention To Detail; Strong Written And Verbal Communication Across Design, Validation, And Documentation Workflows.
- Security: C/C++ Memory Safety, Fuzzing (Oss-Fuzz Methodology), Vulnerability Triage And Patch Authoring, Benchmark Construction Against Cybergym / Arvo / Oss-Fuzz Corpora.
- Infrastructure: Docker, Kubernetes (Aws Eks), Terraform, Github Actions, Fastapi, Flask, Postgresql, Redis, Next.Js/Typescript.
Languages
Education
Ankara Üniversitesi
B.Sc. · Software Engineering · Ankara, Turkey
Classes Préparatoires aux Grandes Ecoles (CPGE)
Mathematics & Applied Reasoning
Lycée d'Excellence 2
High School Diploma · Science · Nouakchott, Mauritania
Statistics
Experience
Global experience
Expertise
Qualifications
Profile
Frequently asked questions
Have questions? Find more information here.
Mohamed is based in Çankaya, Turkey.
Mohamed speaks the following languages: Arabic (Native), French (Native), English (Advanced), Turkish (Advanced).
Mohamed has at least 5 years of experience. During this time, Mohamed has worked in at least 4 different roles and for 4 different companies. The average length of individual experience is 1 year and 3 months. Note that Mohamed may not have shared all experience and actually has more experience.
Based on recent experience, Mohamed would be well-suited for roles such as: RL Environment & Benchmark Author, Freelance ML Systems Engineer, Full-Stack & Data Contributor.
Mohamed's most recent position is RL Environment & Benchmark Author at AfterQuery.
In recent years, Mohamed has worked for AfterQuery, Afterquery, Various, and Mattel S.A.
Mohamed is most experienced in industries like Information Technology and Manufacturing.
Mohamed is most experienced in business areas like Information Technology and Product Development.
Mohamed holds a Bachelor in Software Engineering from Ankara Üniversitesi.
Mohamed is immediately available for suitable projects.
Daily rate distribution
The rates shown represent the typical market range for freelancers in this position based on recent contracts on our platform.
Average rates for similar positions
Rates are based on recent contracts and do not include FRATCH margin.
The average daily rate is the mean of all daily rates from recent contracts of comparable freelancers on our platform.
The median daily rate is the middle value of all daily rates — half of comparable freelancers charge less, half charge more. Unlike the average, it is barely affected by outliers.
Calculated based on our freelancers’ daily rates as of 16 Oct 2026. Actual rates may vary depending on seniority level, experience, skill specialization, project complexity, and engagement length.
Similar freelancers
Discover other experts with similar qualifications and experience
Stylianos I.
Kunter K.
Anton K.
Zakaria A.
Nurbüke T.
Kaan D.
Mevlüt Y.
Majid K.
Tim R.
Yahya S.
Experts recently working on similar projects
Freelancers with hands-on experience in comparable project as a RL Environment & Benchmark Author
Sebastian N.
Martin G.
Jens H.
Mbah S.
Osman T.
Achille C.
Dieter M.
Stephan H.
Yannide D.
Atef A.
Nearby freelancers
Professionals working in or nearby Çankaya, Turkey
Alperen Ç.
Ozge S.
Haydar V.
Ezgi A.
Hayati T.
Onur B.
Hedi E.
Mesut K.
Taha S.
Ömer M.
