// SECTION PROGRESS
01 / 09
SYSTEM ONLINE · OPEN TO ROLES

Vishal
Singh

Vishal Singh

I work at the intersection of machine learning and computational science, building physics-informed models, neural operators, and LLM-based systems to model and optimize complex phenomena across healthcare, engineering, and intelligent systems, bridging research and real-world deployment.

Download CV Research See Projects
Code is our instrument --- reality, the dataset
f : ℝⁿ → ℝᵐ
FORWARD PASS

"Decisions Determine Destiny"

02 / About

A mechanical engineer turned ML researcher.

I started with continuum mechanics and ended up training neural networks to solve PDEs. The bridge between physics and data is where I'm most at home.

B.E. in Mechanical Engineering from Thapar Institute of Engineering and Technology (CGPA 9.00/10), now Scientific ML Intern at NMCAD Lab, IISc Bengaluru. My research applies deep learning to aerospace, biomedical, and optimization problems — and I'm equally at home pushing a paper through review or shipping a Streamlit app for clinicians.

Scientific ML Physics-Informed NNs Neural Operators ML for Healthcare ML for Finance Optimization Generative AI
UNIT.CELL· 0×A1
PINN.FIELD· u(x,t)
Tech Stack
Languages
PythonJuliaMATLABLaTeXSQL
LLM Tools & Frameworks
LangChainDSPyUnslothHugging FacePhoenix ArizeMLflowLangfuse
Frameworks
PyTorchTensorFlowKerasOpenCVStreamlit
Developer Tools
GitVS CodeGoogle ColabJupyter NotebookDocker
Libraries
PandasNumPyScikit-LearnMatplotlibSeabornPySINDy
Softwares
PowerBIMicrosoft ExcelCreo ParametricAnsysAutoCAD
03 / Education

Foundation in rigor and curiosity.

Mechanical engineering taught me to model the physical world; the rest came from staying curious about how data could do it better.

KNOWLEDGE.MODEL· CGPA 9.00

B.E. — Mechanical Engineering

2021 — 2025

Thapar Institute of Engineering & Technology, Patiala

CGPA 9.00 / 10 · Top 10% of department
Academic Extracurricular Coursework
  • Graduated with B.E. in Mechanical Engineering — CGPA 9.00/10
  • Top 10% of department; multiple merit scholarships
  • Final-year capstone on a semi-autonomous water-cleaning robot
  • NCC Air Wing Cadet (Nov 2022 — Oct 2023)
  • Core team — SOMIE; organized MATLAB workshops 2022–2024
  • Graphic Designer — Thapar Movie Club (5,000+ students)
  • Student member — ASME
  • Core team — Mechatronics & Robotics Club
Artificial Intelligence
Advanced Calculus
Linear Algebra
Complex Analysis
Differential Equations
Fourier Transform
Laplace Transform
Manufacturing
Strength of Materials
Theory of Machines
Fluid Mechanics
Optimization Methods
Numerical Analysis
Material Science
Mechatronic Systems
Computer-Aided Design & Analysis
Thermodynamics
Industrial Engineering
Automobile Engineering
Vehicle Dynamics
Machine Design
Graph Theory
Fluid Machines
Heat Transfer
Dynamics & Vibrations
Internet of Things
Machine Learning
Operations Management
Deep Learning
Computer Vision
Julia for Machine Learning
AI for Science and Engineering
LLM Engineering
Introduction to Software Engineering
04 / Experience

Six missions across research labs and industry.

From CFD pipelines at IIT Delhi to surrogate models at JuliaHub — pick a mission to see the brief.

IISc Bengaluru
SciML Intern
Apr 2025 — Now
JuliaHub
ML Intern
Jan — Mar 2025
Samsung R&D
PRISM Intern
Oct 2024 — Apr 2025
TIET / IIITDM
UG Researcher
Feb 2024 — Jun 2025
École Polytechnique MTL
MITACS Intern
Jun — Aug 2024
IIT Delhi
DL Intern
Jun 2023 — Jun 2024
NMCAD LAB · IISc BENGALURU
Scientific ML Intern
APR 2025 — PRESENT

Research on auxetic metamaterials, fatigue crack growth prediction, and Industry 4.0 applications — using physics-informed neural networks and hybrid FEM-PINN solvers.

3
Journals
2
Conferences
FEM+PINN
Method
  • Built FEM-PINN hybrid solver for fatigue crack growth analysis
  • 3 journal papers + 2 conference papers published or submitted
  • Presented work at ICESM 2025
JULIAHUB · BENGALURU
Machine Learning Intern
JAN 2025 — MAR 2025

Enhanced JuliaSimSurrogates for GUI deployment and optimized the Surrogatize.jl testing pipeline, cutting test time by 50%.

50%
Test Time Cut
GUI
Deployment
Julia
Stack
  • Cut Surrogatize.jl test runtime from 3 hours down to 1.5 hours
  • Integrated surrogate tuning into cloud workflows
  • Polished JuliaSimSurrogates UI for GUI deployment
SAMSUNG R&D · REMOTE
Samsung PRISM Intern
OCT 2024 — APR 2025

Built a video-generation framework using diffusion models, working under Samsung's Visual Intelligence team.

Diffusion
Method
UNet
Arch
Video
Generation
  • Designed temporal consistency module for video diffusion
  • Implemented lightweight UNet for faster generation
TIET & IIITDM KANCHEEPURAM
Undergraduate Research Assistant
FEB 2024 — JUN 2025

Explored PINNs, Neural Operators, KANs and PI-KANs applied to aerospace, biomedical and computational mechanics problems.

PINN
Framework
KAN
Arch
Patent
Filed
  • First-author paper on PINNs for biomedical analysis
  • Trained KANs on multi-physics simulation problems
  • Received internal R&D grant for healthcare AI
ÉCOLE POLYTECHNIQUE DE MONTRÉAL · CANADA
MITACS Globalink Research Intern
JUN 2024 — AUG 2024

Fully funded research developing FNO, DeepONet, and PINO models for porous media flow simulation.

FNO
Arch
3D
Darcy Flow
MITACS
Fellowship
  • Trained FNO and PINO on 3D Darcy flow datasets
  • Custom collocation loss for PINO training
  • Published internal report on permeability prediction
SMICR GROUP · IIT DELHI
Deep Learning Intern
JUN 2023 — JUN 2024

Predicted boundary-layer thickness on LPT blades using FNO at 91% accuracy and built ANN models for aerodynamic coefficients.

91%
FNO Accuracy
ANN
Aero Model
CFD
Pipeline
  • 91% accuracy with FNO on boundary-layer thickness
  • Solver-free aerodynamic prediction pipeline
  • Integrated CFD data into ML-ready format
06 / Research

Papers, patents and book chapters.

Mostly on physics-informed learning, neural operators and metamaterial design. A complete list lives on Google Scholar — selected highlights below.

FIELD · u(x,t)· ∇·u = 0

Physics-Informed Neural Network for a 1D Solid Mechanics Problem

Singh, V., Harursampath, D., et al.

Modelling, 2024 — 5, 1532–1549

DOIPublished

Detection of Structured False Data Injection in RL-Controlled Artificial Pancreas Systems

Vishal Singh, et al.

Submitted to IEEE Transactions on Industrial Informatics

Under Review

Neural Optimization Machine for Auxetic Metamaterial Design

Vishal Singh, et al.

Submitted to Journal of Computational Science

Under Review

Neural Optimization of Anti-Tetrachiral Auxetic Metastructures

Vishal Singh, et al. — Ongoing manuscript

In Progress

KAN-Based Prognostics for Remaining Useful Life of Aircraft Engines

Vishal Singh, et al. — Ongoing manuscript

In Progress

PINN for Glucose-Insulin Dynamics in Type-1 Diabetes

Vishal Singh, et al.

AICOMAS 2025 · Paris

Physics-Enhanced NN for Re-Entrant Auxetic Structures

Vishal Singh, et al.

MSML 2025 · Naples

Pseudo-Artificial Pancreas — AI-Based Insulin Dosage Recommendation

A low-cost decision-support system for Type-1 diabetes management.

Patent Filed

Optimal Auxetic Metamaterial Structures — AI-Based Design Framework

Vishal Singh

I-4AM 2026, IISc / Springer (forthcoming)

Book Chapter

Piezoelectric Auxetic Energy Harvesting Devices

Vishal Singh

I-4AM 2026, IISc / Springer (forthcoming)

Book Chapter
05 / Projects

Things I built and actually shipped.

Mostly research-adjacent — surrogate models, ML web apps, a robot. The kind of code that touches real users or real physics.

STREAMLIT · LR

Breast Cancer Prediction Web App

Streamlit app reaching 97% accuracy on diagnostic data — logistic regression, radar-chart visualisation, modular deployment.

Pythonscikit-learnStreamlit
Live App
RAG · CHATBOT

Jupiter Chatbot

A retrieval-augmented assistant for Jupiter.Money that answers questions grounded in their official services and policies.

RAGLangChainStreamlit
Live App
CAPSTONE · IOT

Semi-Autonomous Water Cleaning Robot

ESP32 + BTS7960 river-cleaning robot with full ANSYS simulations of the propulsion system. Final-year capstone.

ESP32AnsysCAD
Project Site
FNO · OPERATOR

FNO — Porous Media Flow

Fourier Neural Operator capturing spatiotemporal dynamics of fluid flow across simulation conditions.

PyTorchFNOCFD
Internal Report
GPR · FINANCE

Apple Stock Prediction · GPR

Gaussian Process Regression with uncertainty quantification, custom feature engineering and hyperparameter tuning.

GPRscikit-learnPandas
GitHub
PINN · BIOMED

Glucose-Insulin PINN

Physics-informed network modeling Type-1 diabetes glucose dynamics — basis for the patent-pending pseudo-pancreas.

PINNPyTorchHealthcare
AICOMAS 2025
07 / Achievements

A handful of well-earned milestones.

Scholarships, fellowships, departmental ranks — markers along the way that keep the work honest.

CONSTELLATION· 10 / 10

Merit Scholarship III · 2023–24

INR 1,59,000 for top-10% rank in Mechanical Engineering at TIET.

MITACS Globalink Fellowship

Fully-funded summer research at École Polytechnique de Montréal, Canada.

Merit Scholarship III · 2022–23

INR 1,39,000 for top-10% rank in Mechanical Engineering at TIET.

3rd Rank — Mech Engg · 2022–23

Department-wide academic excellence certificate plus award.

Internal R&D Grant

Funded the early healthcare AI work that became the patent filing.

Multiple Paper Acceptances

MSML 2025, AICOMAS 2025, I-4AM 2026 — across journals and conferences.

08 / Credentials

Verified courses and formal stamps.

Coursework that closed gaps — Coursera, NPTEL, official issuer programs.

Machine Learning Specialization
Coursera · DeepLearning.AI
Deep Learning Specialization
Coursera · Stanford
TensorFlow Developer
Coursera
Python for Data Science
NPTEL
Julia Programming
JuliaAcademy
Foundations of Quantum
Onramp · IBM
Generative AI with LLMs
Coursera · DeepLearning.AI
Computer Vision Basics
Coursera
SIGNED · VERIFIED· 8 / 8
09 / Contact

Channel is open.

Drop a signal — research collaborations, internships, or anything in the AI-for-science neighborhood.

© 2026 Vishal Singh · Built with and code · v 3.0