Parallel Algorithms and HPC Researcher (Research Coding Tasks)
Listing checked October 30, 2026 · pay as published by AfterQuery
Key details
- Pay
- $80/hr
- Location
- Remote
- Job type
- contract
- Posted
- October 29, 2026
Overview
AfterQuery turns peer-reviewed work into evaluation tasks that frontier AI agents must solve, and this role puts your own parallel algorithms research at the center of that pipeline. You take a recent HPC paper with released code, strip out the part that carries the paper's core contribution, and rebuild the surrounding harness so an agent has to recreate that contribution and match the authors' numbers on their own benchmarks. Each task pays $200 once approved and typically takes a few hours. The arrangement is fully remote and asynchronous, so you pick when the work happens.
What You'll Do9
- Choose a recent paper with public code, preferably one you wrote, and identify the piece that holds the main technical contribution.
- Remove that core implementation and package everything around it into a task an AI agent can attempt from a written statement.
- Write the task statement so the goal, inputs and expected behavior are unambiguous without revealing the solution.
- Select held-out workloads for evaluation and set pass thresholds for each metric the task reports.
- Build a hidden reference solution that meets the bar, plus a second solution that demonstrates the bar is beatable.
- Verify that the packaged task reproduces the paper's benchmark conditions and runs from a clean setup.
- Assess agent submissions against your thresholds and record where attempts succeed or fall short.
- Document the task well enough that another reviewer can follow your evaluation logic.
- Iterate on thresholds and workloads when a task proves too easy, too hard or ambiguous.
Requirements7
- Published research in parallel algorithms or HPC at venues such as SC, PPoPP, IPDPS, SPAA or ICS, or comparable outlets.
- Code released alongside at least one of your papers, with a GitHub repository or artifact link you can share.
- PhD in computer science or a related field, either completed or in progress.
- Hands-on depth in MPI, OpenMP, sparse linear algebra, parallel graph algorithms or GPU solvers.
- Track record of first-author papers that report measured gains over a baseline.
- Comfort with Docker and benchmark setups another person can reproduce.
- Ability to scope a research problem into a bounded task with clear pass criteria.
Who Should Apply
Researchers and research engineers who publish in parallel algorithms or HPC and ship working code with each paper. You have run your own benchmarks, know what a reproducible artifact looks like, and can judge whether a solution actually beats a baseline.
Salary Insight
Each approved task pays $200, and the listing estimates a few hours of work per task, which works out to roughly $80 per hour at that pace. Rates at this level match PhD-trained HPC specialists producing evaluation-grade artifacts rather than routine coding.
Required skills
How applying works
1. Pre-screen on NearSkill
A quick application on NearSkill checks your fit against this role.
2. Finish on the employer’s site
We hand you to AfterQuery’s official application page to complete the form.
3. Wait for a response
The employer’s hiring team reviews applications and gets back to you.
Similar jobs
Open roles that match this listing's skills and level.
AfterQueryAlgorithms and Theory Researcher (Research Coding Tasks)
AfterQueryGPU Kernel and ML Systems Researcher (Research Coding Tasks)
AfterQueryProgramming Languages Researcher (Research Coding Tasks)
AfterQueryComputer Architecture Researcher (Research Coding Tasks)
AfterQuerySoftware Engineering Researcher (Research Coding Tasks)
AfterQuery