Pod lead - Math (Terminal Bench Science)

Hello,

Greetings from ZettaMine!!!

Freelancing Opportunity: We are hiring a POD Lead Mathematics Expert (Terminal Bench Science) for a short-term Frontier AI Research, Scientific Computing & AI Model Evaluation initiative.

Location: India

Role: POD Lead Mathematics Expert (Terminal Bench Science)

Educational Eligibility: PhD, postdoctoral experience, or equivalent advanced research experience in Mathematics, Applied Mathematics, Statistics, or a related field.

Mode: Remote

Start: Expected next week

Duration: 3 Weeks

Employment Type: Contractor assignment (no medical or paid leave).

Mandatory Eligibility Criteria

Technical Qualification:

• Strong programming skills in Python, Julia, C/C++, MATLAB/Octave, or computer algebra systems.
• Strong proficiency in Linux and terminal environments.
• Hands-on experience in Numerical Methods, Scientific Computing, Optimization, Symbolic Computation, Computational Statistics, or Mathematical Modeling.
• Strong understanding of mathematical proofs, algorithms, convergence, stability, numerical conditioning, and error analysis.
• Experience reviewing mathematical derivations, research outputs, computational models, or scientific code.
• Ability to validate mathematical correctness, numerical accuracy, solver performance, and reproducibility.
• Experience designing or reviewing automated tests, mathematical benchmarks, or computational evaluations.
• Experience mentoring technical teams, reviewing deliverables, and maintaining quality standards.
• Excellent written communication skills for providing precise technical feedback and documentation.

Preferred Qualifications

• Familiarity with NumPy, SciPy, SymPy, mpmath, SageMath, PARI/GP, or Mathematica.
• Experience with FEniCS, Firedrake, PETSc, DifferentialEquations.jl, CVXPY, JuMP, or Pyomo.
• Familiarity with Lean/mathlib, Coq/Rocq, Isabelle, or other proof assistants.
• Experience with Docker, Git, CI systems, and automated testing.
• Experience designing mathematical benchmarks, competition problems, or simulation-based evaluations.
• Prior experience evaluating AI coding agents, mathematical reasoning, or AI training programs.

What You'll Work On

• Lead and mentor a pod of approximately 510 technical trainers developing realistic, terminal-based mathematical tasks for a frontier AI training and evaluation initiative.
• Review tasks end to end, including problem statements, datasets, computational environments, reference solutions, and automated tests.
• Validate tasks involving numerical analysis, optimization, statistics, differential equations, mathematical modeling, computational algebra, and formal verification.
• Ensure tasks require genuine multi-step reasoning rather than simple lookups or single-library calls.
• Verify convergence, stability, numerical conditioning, discretization errors, and appropriate computational precision.
• Review automated graders to ensure they evaluate mathematical properties, residuals, feasibility, optimality, error bounds, and proof correctness.
• Identify mathematical inaccuracies, unrealistic assumptions, hard-coded outputs, and opportunities to bypass intended computations.
• Provide actionable feedback, track revisions, and ensure quality issues are resolved before task approval.
• Onboard and mentor trainers, allocate work, monitor quality and productivity, and resolve technical blockers.
• Maintain technical documentation, task design guidelines, and best practices.

Availability Requirements

• Minimum 40 hours per week.
• At least 4 hours of availability per day.
• Minimum 4 hours of overlap with PST business hours.

Interested candidates: Please share your updated CV and LinkedIn profile at karthik.k@zettamine.com.

Thanks & Regards,

Karthik K

ZettaMine