Suraj Pawar Email & Phone Number

Houston, Tx, Us

London, England, Gb

Responsible for innovation, algorithm development, and productizing deep learning applications for a wide range of problems in the energy industry.

Stillwater, Ok, Us

Led scientific ML research for Earth system modeling, achieving significant contributions and technological advancements.• Proposed an equivariant CNN model for turbulence closure modeling by incorporating physical symmetries as constraints, achieving a 32% reduction in loss and improved correlation compared to a CNN, along with stable deployment. • Integrated Kalman filtering and transformer neural operator to learn closure models for large eddy simulation of turbulent flows, reducing run time by 40% in the a posteriori deployment. • Engineered a novel probabilistic surrogate model using deep-ensembles for turbulent flow reconstruction tasks, improving prediction accuracy by up to 20% and reducing uncertainty by up to 10%. • Implemented a nonintrusive ROM framework using an LSTM network for data-driven autoregressive forecasting of geophysical flows, achieving a 100X computational speedup in the data assimilation loop. • Developed a multi-objective evolutionary strategy algorithm for hyperparameter optimization and automatic pruning of CNNs used in surrogate modeling of geophysical flows, reducing the model’s computational complexity by 50%.

Golden, Co, Us

• Designed a multi-fidelity physics-informed neural network framework for wake modeling of a wind turbine, attaining a maximum relative percentage error for the kinetic energy flux of less than 1%. • Worked with a cross-functional team of engineers and data scientists to integrate the model into NREL’s wind farm optimization toolkit.

Lemont, Il, Us

• Led the development of a distributed reinforcement learning library for scientific simulation on supercomputers, contributing to a 30% computational acceleration in the convergence of steady-state CFD simulation. • Deployed and analyzed the scalability of the Ray library coupled with OpenFoam on 512 nodes with 8 CPUs each.

1. Performed preliminary design of the propulsion system of the autonomous underwater vehicle which can tow a large payload2. Verified and validated RANS Gamma-Re Theta transition model with experimental results for open water performance of ducted propeller KA-4-703. Developed a self-propulsion model for the autonomous underwater vehicle to predict the vehicle drag and to find out various means to reduce the vehicle drag4. Developed surrogate-based design optimization framework for ducted propeller by combining high fidelity Reynolds Averaged Navier Stokes Equation (RANS) numerical model and parametric geometry definition of the propeller in terms of pitch distribution, camber distribution, and airfoil thickness distribution5. Working on automation of simulation workflow from geometry generation to post-processing for ducted propellers in StarCCM+ using user-defined scripts6. Designed and integrated the post-swirl stator behind the propeller to counteract the rolling torque of the propeller

1. Achieved on-time delivery of 94% of outsourced projects components like castings, forging, raw material, fasteners, Hardware items multiple projects. Represented the department in a special task force formed to expedite the completion of export projects3. Started maintaining the project progress report to monitor the progress of project completion and reported bottlenecks to the management. Developed the base plan in Microsoft Project for long-lead items like Auxiliaries, Castings, Forgings and Raw material and ensured its compliance to meet the on-time delivery 5. Accomplished the cost saving of 0.47% of the total budget allocated to the procurement of outsourced components by coordinating with the production planning and design department6. Monitored, analyzed, and reported the financial performance (actual vs. budget) of simultaneous high-volume projects on a monthly basis to meet cost/budget limitations and schedules7. Formulated the proposal for bulk procurement of bought out items taking into account the future project requirement with acceptance and currently in the implementation phase7. Assisted Project Management department in Vendor proposal and their approval for NTPC Khargone project. Coordinated with vendors for drawing approvals from the customers of the NTPC project. 8. Successfully completed 3 months of Power Plant Familiarization training at Power Training Institute, L&T LTD. Underwent hands-on training for various manufacturing processes like Milling, Drilling, Lathe Machining, welding, fit-up, and assembly

Mumbai, Maharastra, In

1. Developed a 2-Dimensional Finite Difference Transient Thermal model to predict the temperature distribution in the cold-rolled aluminum coil during the annealing process which will be used by process engineers at the plant.2. Designed an experiment to determine the variation of thermal contact conductance of aluminum coil with applied compressible pressure using a four-ball tester machine.3. Compared the effect of critical parameters on annealing cycle time and temperature distribution within the aluminum coil.4. Calibrated the Mathematical Model for two annealing furnaces of Hindalco Industries, Taloja Plant, India.5. The mathematical model was developed using the open-source programming language Python. The GUI was developed using PyQT which allowed the user to run multiple cases of annealing cycle to find the suitable annealing cycle for the furnace.

Doctor Of Philosophy - Ph.D, Mechanical Engineering

Master’S Degree, Mechanical Engineering

Bachelor Of Technology (B.Tech.), Mechanical Engineering