10/27/2026 updated
100 % available
Physics Subject Matter Expert | AI Output Validation & Scientific Problem-Solving
Jaipur, India
Only remote
Master of Science (M.Sc.), PhysicsAnalytical ThinkingComputational PhysicsCreative Problem SolvingEngineering PhysicsProblem SolvingPhysicsQuantum
Physics Expertise
Deep knowledge across mechanics, oscillations, electromagnetism, wave propagation, thermodynamics, statistical physics, stochastic dynamics, and quantum information, enabling the design of rigorous and self-contained physics problems.
AI Evaluation & LLM Output Validation
Expertise in LLM response analysis, benchmark problem design, reference-answer validation, rubric development, and reasoning-error diagnosis to assess and improve the scientific accuracy of AI-generated solutions.
Programming & Data Analysis
Proficiency in Python with libraries such as NumPy, SciPy, and SymPy for numerical simulations, symbolic computation, scientific calculations, and data analysis, as well as LaTeX for technical documentation.
Scientific Methods
Application of mathematical modeling, symbolic verification, numerical quadrature, master-equation integration, boundary-condition checks, and asymptotic analysis to solve complex physics problems.
Scientific Software
Hands-on experience with specialized scientific software including Quantum ESPRESSO, LAMMPS, ASE, pymatgen, FEniCSx, and PyBaMM for advanced simulations and computations.
Quantum Information
Work on quantum information problems involving Bell inequalities, CHSH correlations, and entanglement, including analysis of noise and physical constraints limiting achievable correlations.
Benchmark Development for LLMs
Creation of original physics problems and step-by-step reference solutions with task-specific evaluation rubrics covering conceptual correctness, mathematical rigor, and constraint handling to assess LLM capabilities.
Stochastic Dynamics & Statistical Physics
Analysis of thermodynamics and statistical physics problems on entropy, equilibrium populations, detailed balance, and the Boltzmann distribution, linking microscopic models to macroscopic behavior.
Deep knowledge across mechanics, oscillations, electromagnetism, wave propagation, thermodynamics, statistical physics, stochastic dynamics, and quantum information, enabling the design of rigorous and self-contained physics problems.
AI Evaluation & LLM Output Validation
Expertise in LLM response analysis, benchmark problem design, reference-answer validation, rubric development, and reasoning-error diagnosis to assess and improve the scientific accuracy of AI-generated solutions.
Programming & Data Analysis
Proficiency in Python with libraries such as NumPy, SciPy, and SymPy for numerical simulations, symbolic computation, scientific calculations, and data analysis, as well as LaTeX for technical documentation.
Scientific Methods
Application of mathematical modeling, symbolic verification, numerical quadrature, master-equation integration, boundary-condition checks, and asymptotic analysis to solve complex physics problems.
Scientific Software
Hands-on experience with specialized scientific software including Quantum ESPRESSO, LAMMPS, ASE, pymatgen, FEniCSx, and PyBaMM for advanced simulations and computations.
Quantum Information
Work on quantum information problems involving Bell inequalities, CHSH correlations, and entanglement, including analysis of noise and physical constraints limiting achievable correlations.
Benchmark Development for LLMs
Creation of original physics problems and step-by-step reference solutions with task-specific evaluation rubrics covering conceptual correctness, mathematical rigor, and constraint handling to assess LLM capabilities.
Stochastic Dynamics & Statistical Physics
Analysis of thermodynamics and statistical physics problems on entropy, equilibrium populations, detailed balance, and the Boltzmann distribution, linking microscopic models to macroscopic behavior.
Experience
Professional experience0 - 5 years
Freelancer since2025
Languages
EnglishFluent
Project history
Design original, self-contained physics problems testing model selection, mathematical abstraction, constraint handling, and multi-step reasoning in LLMs. Develop reference solutions with explicit assumptions, intermediate derivations, and symbolic transformations. Review AI-generated solutions for factual, mathematical, and scientific correctness, and validate answers through analytical reasoning and Python-based symbolic and numerical checks.
Design original, self-contained physics problems on which large language models fail, targeting reasoning errors, invalid shortcuts, and missed constraints. Write reference solutions with explicit assumptions and step-by-step derivations, verified through Python-based symbolic and numerical checks. Analyze model failures and refine problem difficulty to ensure tasks remain rigorous, unambiguous, and technically correct.
Developed research-oriented physics evaluation tasks with supporting evidence, clearly defined deliverables, and expert reference answers. Built task-specific rubrics for scientific reasoning, evidence use, numerical rigor, and uncertainty. Provided detailed error explanations and constructive feedback, revising instructions and solutions in response to evaluation findings.