Mihir Upadhyay

Mihir Upadhyay

New York University

I'm pursuing my Master's in Data Science at the Center for Data Science at New York University. Currently, I'm working under the supervision of Prof. Umang Bhatt. My research focuses on developing algorithms and trustworthy AI systems for effective human-AI collaboration, particularly in high-stakes settings.

Recently, I interned with Prof. Muhammad Bilal Zafar at the Research Center Trustworthy Data Science and Security, Ruhr University Bochum. There, I studied epistemic diversity and temporal awareness in generative search agents, which deepened my understanding of advanced topic-modeling methods.

Prior to my academic pursuits, I worked in industry as a Product Manager at Flixstock. Under the guidance of Dr. Harinder Keer and Ashish Kumar, I researched AI-based image generation using diffusion models, GANs, and vision–language models to controllably replace real humans in e-commerce imagery with AI-generated human models.

Before Flixstock, I was part of the AI tools team at Standard Chartered, working with Thekkath Sabarinath and Amarendra Rout. My work focused on implementing and integrating AI tools (such as Truera and DataRobot) in the banking sector.

I completed my Bachelor's degree at IIT BHU, where I developed a strong interest in data science and machine learning through friends, Kaggle competitions, hackathons, and competitive programming.

I have a deep love for music and enjoy singing, and occasionally channel my inspiration into writing. I plan to add all my original work here over time.

Publications

When Should We Orchestrate Multiple Agents?

Authors: Umang Bhatt, Sanyam Kapoor, Mihir Upadhyay, Katherine M. Collins, Ilia Sucholutsky, Francesco Quinzan, Adrian Weller, Andrew Gordon Wilson, Muhammad Bilal Zafar.

View

Characterizing Web Search in The Age of Generative AI

Authors: Elisabeth Kirsten, Jost Grosse Perdekamp, Mihir Upadhyay, Krishna P. Gummadi, Muhammad Bilal Zafar.

View

Skills

Python

Advanced programming for ML/AI development

SQL

Database management and data querying

JavaScript

Web development and interactive applications

LangChain

LLM orchestration and agent development

FastAPI

High-performance API development

Apache Spark

Big data processing and distributed computing

PyTorch

Deep learning and neural network development

NLTK

Natural language processing and text analysis

Agentic AI

Multi-agent systems and autonomous AI

Experience

Visiting Research Assistant

University of Cambridge | Sep 2025 – Present | New York, United States (Remote)

  • Led the development of an open-source framework to systematically characterize second- and third-order effects of human-AI interactions on downstream human-human decision-making.
  • Recorded & analysed 10,000+ dyadic multi-player interactions via Prolific using custom Empirica-based experimental pipelines.

Research Intern

Research Center Trustworthy Data Science and Security, Ruhr University | May 2025 – Aug 2025 | Bochum, Germany

  • Systematically analyzed methods for evaluating the diversity of information presented by generative content retrieval agents vs. traditional web search (Under review @ ARR).
  • Curated 6 real-world query datasets and collected search results via various search APIs, analyzing source coverage and temporal patterns.
  • Designed a human evaluation study using Prolific to measure user preferences across Google AI Overviews, Gemini, and GPT-4 pipelines.

Product Manager

Flixstock | Oct 2022 – Sep 2023 | Gurugram, India

  • Designed and deployed e-commerce image generation pipelines using diffusion models, GANs, and image captioning, enabling AI models to replace human models in 20% of the catalog.
  • Directed a team of 8+ engineers to develop and launch three full-stack platforms with REST API endpoints, deployed on AWS.

AI/ML Analyst

Standard Chartered Bank | Jul 2021 – Sep 2022 | Bengaluru, India

  • Evaluated and deployed bias, drift, and performance monitoring tools (Truera, DataRobot) for 10+ forecasting models, benchmarking against traditional ML methods.
  • Led a 5+ member team to deploy tools via OpenShift and Kubernetes (Docker containers), leveraging Kubernetes operators for lifecycle management and Jenkins for CI/CD.

Data Science Intern

Standard Chartered Bank | May 2020 – Jul 2020 | Bengaluru, India

  • Built and tuned ensemble classifiers on a European credit card transaction dataset, improving fraud detection and reducing false positives by 5%.
  • Developed Tableau dashboards for KPIs like transaction volume and response time, reducing reporting time by 4 hours/week.

Teaching

Teaching Assistant

Courant Institute of Mathematical Sciences, NYU | Sep 2025 - Present | New York City, NY, USA

  • Course: Algebra, Trigonometry, and Functions

Teaching Assistant

Center for Data Science, NYU | Jan 2025 - May 2025 | New York City, NY, USA

  • Course: Responsible Data Science

Education

Master of Science in Data Science

New York University | Sep 2024 – Present | New York City, NY, USA

  • Overall GPA: (3.93/4.0)
  • 1st paper under review
  • 2nd paper under review
  • SVSAES Scholarship for Academic Excellence
  • Member Graduate Community Building group

Bachelor of Technology in Chemical Engineering

Indian Institute of Technology BHU | Jul 2017 – May 2021 | Varanasi, India

  • Overall GPA: 8.68/10.0
  • UG project on AI methods in Computational Fluid Dynamics
  • IIT BHU Merit Scholarship for the academic years 2017-2021
  • Secretary, Western Music Club

Recommendations

Research Papers

Books

Fiction
  • The Family Upstairs by Lisa Jewell
Non-Fiction
  • Coming soon....

Music

  • Coming soon....

News

Started Capstone project with IBM Research! Sep 2025
Started summer internship as research intern at RC Trust (Bochum, Germany) May 2025