Headline

Lead Data Scientist

About

Machine Learning | AI for Engineering | Filtration Optimization I have a strong background in machine learning, deep learning, and AI-driven optimization for filtration systems. My expertise lies in applying data-driven insights and cloud solutions to enhance product design, improve efficiency, and drive innovation in industrial applications. Key Highlights: 🔹 SmartFE – AI-Optimized Filtration: Built ML-driven workflows to optimize filter design, reducing development cycles from months to weeks. Deployed on Azure with a Streamlit interface for real-time AI recommendations. 🔹 autoDHC – ML for Filter Media Performance: Predicted pressure drop & dust holding capacity (DHC) using Darcy’s Law & regression models, integrating ANSYS CFD simulations and MLflow tracking. 🔹 AI Segmentation of Fibers (CT Scans): Developed a 3D U-Net deep learning model to segment mixed fibers in filter media CT scans, deployed as an Azure-based batch-processing system. 🔹 autoGrid DP – Predicting Pressure Drop: Created an ML tool to estimate pressure drop in irregular filters, extracting C1 & C2 resistivity coefficients via clustering & regression for CFD simulations. I am passionate about AI-driven engineering solutions, constantly exploring ways to improve predictions, optimize designs, and enhance performance. Looking forward to connecting and collaborating on cutting-edge innovations!

Location

Bengaluru

AI stack

aistudio

hugging face

langchain

langgraph

ollama

openai

Work experience

Tech Lead special at Societe Generale Global Solution Centre, May 2025 - Present · 1 yr 4 mos, Bengaluru, Karnataka, India

Senior Filter Media Simulation engineer at MANN And HUMMEL Filter Private Limited, Aug 2019 - Jun 2024 · 4 yrs 11 mos, Bengaluru Area, India, 1.AI-Optimized Smart Filtration (SmartFE) : a.Developed an AI-driven workflow to optimize filter designs using machine learning. b.Integrated with PVS for automated data retrieval and validation. c.Deployed on Azure with Docker for scalability. d.Built a Streamlit interface for real-time AI recommendations. e.Cut design cycles from months to weeks. 2.autoDHC – ML for Filter Media Optimization a.Predicted pressure drop & dust holding capacity (DHC) with ML models. b.Applied Darcy’s Law & regression for airflow resistance analysis. c.Integrated MLflow for tracking and deployed via REST API. d.Enabled ANSYS CFD simulations to validate predictions. 3.AI Segmentation of Fibers in Filter Media (CT Scans) a.Built a 3D U-Net model (TensorFlow) for fiber segmentation. b.Applied clustering & morphological operations for accuracy. c.Deployed batch processing on Azure Custom Models. d.Automated fiber analysis, improving filter R&D. 4.autoGrid DP – Predicting Pressure Drop for Irregular Filters a.Developed an ML tool for airflow resistance in complex filters. b.Extracted C1 & C2 resistivity coefficients using clustering & regression. c.Integrated into CFD simulations for accurate pressure drop predictions.

multiphysics cae engineer at ATOA SCIENTIFIC TECHNOLOGIES, Jul 2018 - Jul 2019 · 1 yr 1 mo, Greater Bengaluru Area, 1.Multiphysics simualtion for Vehicle battery peformance analysis and enhancements . 2. Multiphysics simuation of Non invasive process to detect blood clot in brain using 3D human head CAD. 3. Multiphysics simulation of Exhaust gas recirculation to generate electricity and charge the battery using Thermo electric generators in vehicles .

Graduate Research Assistant at IFSTTAR, Apr 2017 - May 2018 · 1 yr 2 mos, Nantes, Pays de la Loire, France, Developing a numerical model of a section of XLPE insulator in COMSOL Multiphysics software, with water trees represented as voids. Defining material properties and interaction constraints in COMSOL, including those for XLPE and the voids. Constructing a finite element model and generating a suitable mesh to minimize the impact of external parameters. Performing Finite Element Analysis (FEA) by applying varying loads in terms of voltage and analyzing the displacement of voids. Comparing theoretical and analytical results with FEA results to validate the model. Using level set techniques to locate and track the evolution of voids within the insulator.

Research Intern at IFSTTAR, Mar 2016 - Feb 2017 · 1 yr, Greater Nantes Metropolitan Area, Numerical Model Development in python: Develop a numerical model for carbon cord stitching. Define the size and placement of holes on the trailing edge using optimization techniques. Material Properties and Contact Constraints in Python: Define user-defined material properties. Establish contact constraints using Python scripts. Finite Element Model Construction and Analysis: Generate a suitable mesh for the Finite Element Model (FEM) to minimize the impact of external parameters. Perform FEM analysis of the structure with various parameters using commercial FEA solvers such as Abaqus and COMSOL. Robustness and Manufacturing Feasibility Testing: Test the robustness and ease of manufacturing feasibility of the specimen. Conduct both quasi-static and dynamic analysis to ensure the reliability of the model. The project aims to improve the bonding strength of the structure by analyzing the influence of different parameters through a combination of Python for optimization,material properties and constraints definition, and commercial FEA solvers for comprehensive analysis.

Education

Master's in Mechanical Engineering- Applied Mechanics, Compmutational Mechanics, Centrale Nantes

Bachelors, Mechanical Engineering, Dr. Ambedkar Institute of Technology