AI Scientist working on forecasting, optimization, and autonomous decision-making.

Machine Learning • Deep Learning • Reinforcement Learning (PPO, TD3, SAC, DQN) • Physics-Informed ML • Time-series forecasting • Computer Vision • Scalable pipelines (AWS, Spark, HPC)

📍 Los Alamos, NM, USA
singhshubhendu29@gmail.com
Download CV (PDF) Google Scholar GitHub LinkedIn Contact
Shubhendu Kumar Singh profile photo
Highlights
  • AI Researcher at Los Alamos National Laboratory (03/2024–Present)
  • Global streamflow modeling across 20,000+ river basins
  • Multi-action / multi-agent RL for reservoir control under climate uncertainty
  • 33% accuracy improvement via physics-informed LSTM (LANL internship)
  • 598 citations • h-index 9 • i10-index 9 (Google Scholar)

Profile

Artificial Intelligence (AI) Scientist with expertise in Machine Learning, Deep Learning, and Reinforcement Learning, delivering production-ready models for forecasting, optimization, and autonomous decision-making. Demonstrated impact across energy, manufacturing, climate, and fault diagnostics, leveraging Python, PyTorch/TensorFlow, AWS, and large-scale data pipelines.

Skills

AI / ML

Machine LearningDeep LearningReinforcement Learning PPOTD3SACDQN Physics-Informed MLOptimizationControl Systems LLMsFoundation ModelsNLP Time-series ForecastingSpatiotemporal Modeling Computer VisionOpenCV

Tools / Platforms

PythonMATLAB TensorFlowKerasPyTorch Stable-Baselines3Hugging Face TransformersOpenAI Gym Apache SparkSQL AWSSageMakerBedrock HPC ClustersGoogle Colab

Applied Domains

Fault Diagnostics & PrognosticsAutonomous Systems ManufacturingAdditive Manufacturing Energy SystemsClimate & Hydrological Modeling Structural Health MonitoringEnvironmental Modeling

What I build

  • Autonomous control policies under constraints and uncertainty
  • Physics-aware forecasting models for reliability in extremes
  • Production-ready ML pipelines on cloud/HPC
  • Decision-support tools for climate, energy, and engineering systems

Professional Experience

03/2024 – Present
Artificial Intelligence (AI) Researcher
Los Alamos National Laboratory — Los Alamos, NM, USA
  • Designed and implemented Deep RL (PPO, TD3, SAC) multi-action and multi-agent reservoir control policies for adaptive releases under climate uncertainty.
  • Contributed to global-scale deep learning for streamflow prediction across 20,000+ river basins for climate-impact assessment and large-area forecasting.
  • Developed CNN-LSTM spatiotemporal models for wildfire risk forecasting and mosquito activity prediction for early warning and decision support.
  • Developed and evaluated RL-based mesh generation for CAD models to reduce manual tuning and improve simulation efficiency.
  • Applied LLMs, foundation models, physics-informed ML, AWS, Spark, and HPC workflows to accelerate scientific modeling pipelines and deployment.
08/2020 – 12/2023
Research Assistant
Clemson University — Greenville, SC, USA
  • Developed hybrid Physics-Informed ML architectures for diagnostics, prognostics, and manufacturing systems with improved robustness under sparse/noisy sensors.
  • Led AI model development for the ONR-funded SCOTTY Project using LSTMs, GANs, CNNs, ensemble learning, and RL on multi-sensor maritime data.
  • Applied PPO and DQN for real-time fault mitigation in engine systems, demonstrating closed-loop AI control feasibility in safety-critical settings.
  • Developed and deployed Physics + LSTM frameworks for energy-efficient CNC grinding (DOE-CESMII), achieving ~16% reduction in specific energy consumption in an industrial testbed (ITAMCO).
05/2022 – 08/2022
Machine Learning Intern
Los Alamos National Laboratory — Los Alamos, NM, USA
  • Developed physics-informed LSTM models for streamflow and flood prediction in the Colorado River Basin using time-series + geospatial data.
  • Improved prediction accuracy by 33% through physics-aware constraints.
08/2018 – 08/2020
Research Assistant
University at Buffalo — Buffalo, NY, USA
  • DARPA PAI-funded work on physics-infused ML for predictive modeling of complex dynamical systems.
  • Built Physics-Infused LSTM and CNN-LSTM frameworks for UAVs, SHM, and biological systems under limited/noisy data.
10/2016 – 04/2017
Engineer
TAFE Ltd. — Chennai, India
  • Diagnosed powertrain and drivetrain systems in agricultural tractors; built domain expertise later used in AI-based diagnostics/prognostics research.
10/2015 – 07/2016
Engineer
Cheema Boilers Limited — Chandigarh, India
  • Performed thermal design and optimization of boiler pressure parts and piping; foundational physics-driven modeling experience.

Education

Ph.D. Automotive Engineering

Clemson University — Greenville, SC, USA
08/2024 (as listed in CV)

M.S. Mechanical Engineering

State University of New York at Buffalo — Buffalo, NY, USA
09/2019

B.E. Mechanical Engineering

Birla Institute of Technology — Ranchi, India
06/2014

Awards

1st Place — ASME CIE Hackathon (In-Process Data Mining Challenge), 2020

Developed physics-infused deep learning models including a Physics-Informed LSTM for melt pool size prediction and a cGAN for melt pool shape prediction in metal additive manufacturing.

Publications

Google Scholar
Citations: 575 • h-index: 9 • i10-index: 9
View Scholar Profile

Selected papers (A few from from Google Scholar)

  • Singh, S.K., et al. (2024). Hybrid physics-infused 1D-CNN based deep learning framework for diesel engine fault diagnostics. Neural Computing and Applications.
  • Behjat, A., Rai, R., Chowdhury, S., & Singh, S.K. (2019). PI-LSTM: Physics-infused LSTM network. IEEE ICMLA.
  • Kumar Singh, S., et al. (2024). Deep Learning in Computational Design Synthesis: A Comprehensive Review. JCIS in Engineering.
  • Yang, R., Singh, S.K., et al. (2020). CNN-LSTM architecture for computer vision-based modal frequency detection. MSSP.
  • Nguyen, R., Singh, S.K., & Rai, R. (2023). Physics-infused fuzzy GAN for robust failure prognosis. MSSP.
  • Khawale, R.P., et al. (2024). Digital twin-enabled autonomous fault mitigation in diesel engines: experimental validation. Control Engineering Practice.
  • Singh, S.K., et al. (2024). Reinforcement Learning Meets LSTMs for Optimal Reservoir Control. AGU Fall Meeting Abstract.

Funding & Proposals

Autonomous Integrated Tractor and Spraying System

Funding Agency: South Carolina Department of Agriculture • Amount: $120,000 • Role: Doctoral Researcher & Proposal Developer

Conceptualized, developed technical framework, collaborated with stakeholders, and pitched proposal resulting in successful funding.

Professional Affiliations & Services

Reviewer / Member / Leadership

  • Reviewer — Mechanical Systems and Signal Processing (Elsevier)
  • Reviewer — Computers & Geosciences (Elsevier)
  • Reviewer — Neurocomputing
  • Reviewer — Expert Systems with Applications
  • Reviewer — Structural Health Monitoring (Sage)
  • Reviewer — Journal of Nondestructive Evaluation (Springer)
  • Reviewer — Big Data & Cognitive Computing (MDPI)
  • Reviewer — Journal of Marine Science & Engineering (MDPI)
  • Reviewer — ASME IDETC-CIE
  • Reviewer — SciPy Conference
  • Reviewer — Springer SADHNA
  • Member — PHM Society
  • Member — IEEE
  • Treasurer — Graduate Student Association (UB), Sep 2018–Aug 2020

Invited Talks / Presentations

  • Speaker — IEEE ICMLA (Dec 16, 2019), Boca Raton, FL, USA
  • Invited Speaker — “Physics-Guided Machine Learning” webinar (Aug 1, 2021), Online
  • Guest Speaker — CU-ICAR Lecture Series (Mar 27, 2024), Online
  • Presenter — AGU Fall Meeting (Dec 12, 2024), Washington, D.C., USA
  • Invited Speaker — Los Alamos National Laboratory (Jun 23, 2025), Los Alamos, NM, USA
  • Speaker — HydroML 2025 (May 29, 2025), Lake Arrowhead, CA, USA
  • Presenter — New Mexico Energy Initiatives 2025 Symposium (Nov 18, 2025), Socorro, NM, USA
  • Speaker & Session Organizer — AI in Earth Sciences Workshop 2026 (Mar 23, 2026), Santa Fe, NM, USA

Teaching & Mentoring

Teaching Assistant

  • AuE 4930/6930: Introduction to Data Science, ML, & Search Algorithms — Clemson (Jan–May 2022)
  • MAE 494 Design Projects — University at Buffalo (Jan–May 2020)
  • MAE 451 Design Process & Methods — University at Buffalo (Jan–May 2020)

Mentoring

  • Mentored a master’s thesis (ML & blockchain) — Clemson (Jul–Dec 2021)
  • Mentored a Ph.D. student (ML & hydrology) — LANL (Jun 2024–Nov 2025)

Contact

Email: singhshubhendu29@gmail.com

For collaborations, research roles, consulting, and applied ML projects.

Google Scholar GitHub LinkedIn