Jigyasa Gaurav

Hi, I'm Jigyasa

I am a PhD student in the Department of Computer Science & Engineering at the University of Minnesota, Twin Cities, advised by Prof. Aryan Deshwal.

My research focuses on uncertainty-aware Bayesian optimization algorithms for discovering optimal designs in large offline design spaces — with applications in molecule discovery, protein engineering, and materials design.

My background spans pure mathematics (combinatorics, graph theory, differential topology) at IIT Bombay and IISc Bangalore, and two years engineering deep learning systems in computer vision and NLP at V-Labs. Outside research: avid badminton player, former theater member at IISc's Rangmanch, and enthusiastic teacher.

Research Interests
  • Bayesian Optimization & Black-Box Function Optimization
  • Uncertainty Quantification in Machine Learning
  • Computer Vision & Natural Language Processing
  • AI for Scientific Discovery — drug design, materials science, protein engineering
  • Combinatorics & Graph Theory
Latest News
Apr 2026 Presented a poster at the Research Computing Exhibition 2026 and won 2nd prize.
Feb 2026 Attended the WE Disrupt Conference 2026.
Dec 2025 Volunteered at DevFest MN 2025.
Nov 2025 Attended the Women Innovators Conference 2025.
Sep 2025 Started as CSGSA Event Coordinator at the CS Graduate Students Association, University of Minnesota.
Sep 2025 Started as Teaching Assistant for CSCI 5527 Advanced Machine Learning at the University of Minnesota.
Oct 2025 Volunteered at the Twin Cities Marathon 2025.
Aug 2025 Presented a poster and volunteered at the AI4Science Workshop 2025.
Jan 2025 Joined Prof. Aryan Deshwal's lab to work on offline black-box function optimization.
Jan 2025 Started PhD in Computer Science & Engineering at the University of Minnesota, Twin Cities.
Jul 2024 Concluded two years at V-Labs as an AI Developer in the Computer Vision & NLP team.
I'm always happy to chat about research, collaboration, or anything at the intersection of mathematics and machine learning. Feel free to reach out at gaura012@umn.edu.

Experience

Work

AI Developer

V-Labs (Vijna-Labs) · Computer Vision & NLP Team
Jun 2022 — Jul 2024

Document Extraction using Transformers (DExTr) LLMsTransformersRAG

Built a transformer-based enterprise extraction platform using ChatGPT, LLAMA, and Mistral. Improved context-awareness, reduced hallucinations, and automated insurance workflows (NAS, WC, Loss Run).

Confidence Score Analysis XGBoostCatBoostCalibration

Investigated failure modes of high-accuracy DL models via feature engineering and binary classifiers to identify when predictions are unreliable despite high overall accuracy.

First Notice of Loss (FNOL) CLIPTIMM

Automated insurance hail-damage assessment from images. Developed end-to-end CLIP pipeline for roof classification, improving accuracy by 13% over prior model.

Visual Document Understanding PyTorchDonutDenseNetEasyOCR

Pipelines for classifying noisy scanned financial documents and key information extraction using transformer and CNN architectures with OCR and object detection.

Education

Ph.D. Computer Science & Engineering

Advisor: Prof. Aryan Deshwal · Jan 2025 – Present

MPhil, Mathematics

Combinatorics & Graph Theory

MS Research, Mathematics

BS Research, Mathematics

Research & Projects

Research

Offline Black-Box Function Optimization

Jan 2025 – Present
Univ. of Minnesota
Prof. Aryan Deshwal · University of Minnesota

Developing uncertainty-aware Bayesian optimization algorithms for discovering optimal designs in large offline design spaces. Applications in molecule/therapeutic discovery, protein engineering, and material design.

Chromatic Number of Hypergraphs

2021 – 2022
Univ. of Maryland
Prof. David Harris · University of Maryland, College Park

Explored algorithmic complexity of computing chromatic numbers; analyzed approximation algorithms for maximum independent sets in geometric graphs. Applications in CV, information retrieval, and network routing.

Extremal Combinatorics: Ramsey Numbers & Sunflower Lemma

IIT Bombay
Prof. Niranjan Balachandran · IIT Bombay

Studied exponential improvements on lower bounds of diagonal multicolour Ramsey numbers and improved bounds for the sunflower lemma via structured vs. pseudo-random approaches. Applications in circuit complexity and cryptography.

Translation Numbers & Quasi-homomorphisms on Automorphisms of the Circle

IISc Bangalore
Prof. Siddhartha Gadgil · IISc Bangalore

Showed translation numbers are conjugacy invariants via quasi-homomorphisms. Studied dynamical invariants on area-preserving diffeomorphisms of the disc from the signature knot invariant.

Quasi-homomorphisms and Dynamics

IISc Bangalore
Prof. Siddhartha Gadgil · IISc Bangalore

Studied dynamical invariants on the group of compactly supported area-preserving diffeomorphisms of the open unit disk, derived from the signature knot invariant.

Course Projects

Object Detection & Classification

CNN · TensorFlow

YOLO-based object detection on traffic datasets; mood classifier and sign-language digit recognizer using Keras Sequential and Functional APIs.

CNNTensorFlowYOLOKeras

Optimization & Game Theory

Python · NumPy

Markov chain web ranker, minimax adversarial search, BFS shortest path, and a logical inference knowledge-base solver.

PythonNumPyGraph Theory

Election Software

C++

Simulates elections using plurality and STV (Droop quota) algorithms with system testing for robustness and process validation.

C++AlgorithmsVoting Theory

Skills

Machine Learning
Bayesian OptimizationDeep Learning CNNsTransformers LLMsNN Compression
Frameworks
PyTorchTensorFlow NumPyOpenCV XGBoostCLIP
Programming
PythonC C++MATLAB SageLaTeX
Infrastructure
DockerKubernetes Docker SwarmGit
Mathematics
CombinatoricsGraph Theory TopologyOptimization Probability
Languages
EnglishHindi Spanish (Elementary)

Awards & Honors

Research Computing Exhibition 2026 — 2nd Prize

Awarded 2nd prize for poster presentation at the Research Computing Exhibition, University of Minnesota. · Apr 23, 2026

CSIR Junior Research Fellowship (JRF)

Prestigious national fellowship by the Council of Scientific & Industrial Research, India.

INSPIRE Fellowship

Government of India scholarship for top science students, awarded by the Dept. of Science & Technology.

CSIR NET JRF — Qualified

National eligibility exam for lectureship and PhD positions in India. Qualified twice.

GATE — Qualified

Eligible for scholarship and postgraduate admission to premier research institutes in India. Qualified twice.

IISc Open Badminton Championship — Winner

Won individual and team events at the IISc open badminton championship multiple times.

Inter-IIT Badminton Team Member

Represented IIT Bombay at inter-IIT athletic competitions.

Badminton & Table Tennis Medalist

Medalist in several open tournaments at Vijna-Labs.

Professional Service

Conference Reviewer
AISTATS 2026

International Conference on Artificial Intelligence and Statistics.

Event Coordination
CSGSA Event Coordinator

Event Coordinator for the Computer Science Graduate Students Association, University of Minnesota. · Fall 2025 – Spring 2026

Organizer · 'Cryptography' at PRAVEGA

Organised and ran the Cryptography event at PRAVEGA, the annual science & technology festival of the Indian Institute of Science, Bangalore. · 2014

Volunteering
DevFest MN 2025

Volunteered at DevFest MN, a community-organized developer festival. · Dec 6, 2025

Twin Cities Marathon 2025

Volunteered at the Twin Cities Marathon weekend. · Oct 4–5, 2025

AI4Science Workshop 2025

Volunteered at the AI4Science Workshop at the University of Minnesota. · Aug 13–14, 2025

Conferences & Workshops

Research Computing Exhibition 2026 — Poster Presenter & 2nd Prize Winner

Presented a research poster and won 2nd prize. · Apr 23, 2026

SIAM Conference on Uncertainty Quantification 2026

Attended the SIAM UQ conference. · Mar 22–25, 2026

WE Disrupt Conference 2026

Attended the WE Disrupt conference at the University of Minnesota. · Feb 4, 2026

Women Innovators Conference 2025

Attended the Women Innovators conference at the University of Minnesota. · Nov 11, 2025

AI4Science Workshop 2025 — Poster Presenter

Presented a research poster at the AI4Science Workshop, University of Minnesota. · Aug 13–14, 2025

Teaching

Teaching Assistantships
Sep – Dec 2025
CSCI 5527 Advanced Machine Learning University of Minnesota, Twin Cities
Aug – Dec 2020
Advanced ODE Indian Institute of Technology Bombay
Jan – Mar 2020
MA216 Linear Algebra & Calculus Indian Institute of Technology Bombay
Volunteer Instruction
Jan – Sep 2023
Docker in Production & Kubernetes Vijna-Labs

Taught container orchestration to engineering team members.

Oct – Dec 2023
Python Programming Vijna-Labs

Taught Python to annotation team members to build technical proficiency.