Bipen S.
Physics Cohort
Experience
May 2025 - Present
7 monthsPhysics Cohort
Handshake
- Conceived and constructed 20+ intricate, doctoral-level physics prompts in LaTeX, demanding 3-6 layers of multi-step reasoning and using mathematical notation; all prompts featured expert-verifiable and unambiguous solutions.
- Created 20+ structured step-by-step solutions and justifications, reducing ambiguity by >80% and highlighting major reasoning breakdowns in AI responses.
- Pioneered novel AI safety evaluations by translating complex physics principles into reasoning challenges that stumped language models at a rate of 70%+, revealing critical vulnerabilities.
Apr 2025 - Aug 2025
5 monthsPhysics Contributor
Outlier
- Orchestrated physics-based evaluations that exposed key safety issues in AI reasoning, driving fixes for the top three causes of errors.
- Authored 90+ doctoral-level STEM reasoning tasks with 3–10 multi-step layers, achieving a 70%+ model stumping rate and strengthening AI safety.
- Produced 50+ GTFA explanations with 99% accuracy, delivering clear rationales that improved model reasoning assessment and reviewer confidence.
- Revamped evaluation strategies with innovative prompt variations and a standardized scoring system, improving model accuracy by 20% and mentoring new annotators.
- Applied a structured Chain-of-Thought protocol (Assess, Perception, Knowledge Gathering, Advance, Verify, Pivot, Synthesize), embedding deliberate mistakes and self-corrections to strengthen AI training signals and maintain 90%+ compliance with project guidelines.
Feb 2024 - Present
1 year 10 monthsPhysics Specialist (Research)
Sepal AI
- Authored 50+ physics evaluation prompts with step-by-step solutions, enabling precise assessment of AI reasoning capabilities using clearly defined, structured rubrics and accelerated AI performance.
- Built ~10-item rubrics per prompt to assess AI reasoning and highlight improvement areas.
- Diagnosed logical inconsistencies in AI-generated physics solutions using pedagogical methods, advancing model accuracy by refining 50+ outputs weekly, resulting in enhanced educational content quality.
- Spearheaded the development of 10+ rubric frameworks integrating image-based analysis, resulting in enhanced model performance and a significant reduction in inaccurate AI reasoning.
- Conducted extensive experimental work and laboratory research to validate AI model outputs, ensuring accuracy and reliability of physics-based simulations.
Sep 2019 - Apr 2024
4 years 8 monthsAssistant Professor - Physics
Higher Education
- Orchestrated the creation of 3 comprehensive physics curricula for undergraduate and graduate students, incorporating 10+ novel assessment methodologies to evaluate student comprehension of complex concepts.
- Mentored 50+ undergraduate and graduate students on research projects, resulting in 3 peer-reviewed publications and 8 conference presentations at premier universities and national laboratories.
- Constructed 200+ multiple-choice questions (MCQs) aligned with curriculum objectives, enhancing student comprehension by a 15% improvement in average test scores.
Mar 2015 - Mar 2020
5 years 1 monthGeneva, Switzerland
Ph.D. Researcher (CMS Collaboration)
CERN
- Applied physics principles to analyze collider data using Python and C++, boosting data processing efficiency by 35% and contributing to significant advancements in particle physics research at a leading national laboratory.
- Developed Monte Carlo simulations utilizing statistical fitting techniques, refining model accuracy by 25% and reducing prediction errors in particle behavior for precise energy calculations.
- Quantified jet energy correction uncertainties using Iterative Bayesian unfolding, resolving 80% of data–simulation discrepancies.
- Collaborated with 15+ scientists to improve data pipelines and presented results at global conferences.
Summary
PhD physicist with over five years of applied physics experience in AI model evaluation and data annotation for Fortune 500 companies; led the design of complex physics prompts that improved model performance by 20%. Developed rubric frameworks while collaborating with researchers to enhance prompt complexity and domain coverage.
Languages
Hindi
NativeEnglish
AdvancedEducation
Mar 2015 - Mar 2020
Shoolini University
Ph.D. · Experimental High-Energy Physics · Bajhol, India
Aug 2009 - Mar 2012
University of Jammu
M.Sc., Nuclear Physics, Particle Physics, Quantum Mechanics · Physics · Jammu
Certifications & licenses
Physics Specialist (Research)
microl
Need a freelancer? Find your match in seconds.
Try FRATCH GPT More actions
Similar Freelancers
Discover other experts with similar qualifications and experience
- English
- Deutsch
2025 © FRATCH.IO GmbH. All rights reserved.