Sherwin Rajkumar

Contact Information

• Email: [email protected]
• Location: München, Bayern, Deutschland

Summary

Machine Learning Engineer and Data Analyst with a good background in developing and deploying machine learning models to solve real-world and Engineering problems. Proficient in data preprocessing, feature engineering, model selection, and performance evaluation. Adept at leveraging data-driven insights to drive business decisions and enhance processes.

Education

• Master of Science - Computational Mechanics of Materials and Structures (COMMAS)

  • University of Stuttgart, München - Deutschland
  • Graduation Date: May 2023

• Bachelor of Engineering - Mechanical Engineering

  • Anna University, Chennai - India
  • Graduation Date: Apr 2023

Work Experience

IT Consultant - AI and Cloud computing | EVIDEN - An Atos buisness | München, DE | Jun 2023 - Present

• Data and log analytics with Apache Spark, Splunk, ELK and Opensearch.
• MLOps and DevOps.
• Optimization of CAE workflows using AI.
• Investigation of cloud and on-prem LLMs - LoRA, Quantization, LLM Fine-tuning.

Master thesis - Data Analytics and Machine Learning | EDAG Engineering GmbH | München, DE | Nov 2022 - May 2023

• Created a data-driven model to augment dimensioning of brake disks in the early design phase.
• Built a corresponding data-set using vehicle benchmarking platforms such as A2MAC1 and Iceberg.
• Performed EDA and feature selection on the gathered data using Python Matplotlib, Pandas and Seaborn.
• Developed a collection of supervised learning models using scikit-learn and tensorflow.
• Implemented the models as a python based GUI - Tkinter.

Working Student - CAE and Vehicle Safety - Head Impact/Pedestrian protection | EDAG Engineering GmbH | München, DE | May 2022 - Oct 2022

• Developed a complete Finite Element (Explicit dynamics) Head impact scenario with ATD dummies for bicycle helmets according to norms EN 1078 and EN 960.
• Benchmarking of various bicycle helmets based on materials and Head Injury Criterion (HIC).

Internship - CAE and Vehicle Safety - Crash/Body in White (BiW)/Python programming | EDAG Engineering GmbH | München, DE | Nov 2021 - Apr 2022

• Supported the BiW team in building up of FE (Explicit Dynamics) crash models using ANSA, LS-DYNA and Animator4.
• Developed python scripts to automate repetitive routines in ANSA and Animator while pre and post processing FE models.
• End-to-End development of a python script to automate ticket creation of issues in JIRA with data from customer.
• Worked with RESTful services of JIRA and Sharepoint.

Scientific assistant - Reinforcement Learning | Institute für Konstruktionstechnik und Techniches Design | München, DE | Oct 2021 - Apr 2022

• Complete developemt of a Deep-Q learning model to optimize the contour of a press-fitted Shaft-hub joint, to homogenize joint pressure distribution according to norm DIN 7190.
• Material based iterative optimization to achieve the same.
• Created a data-set with optimal contours and trained a supervised-learning DNN model to make predictions of the optimal contour for given design constraints.

Techanical Skills

• Programming Languages: Python, MATLAB, SQL
• Machine Learning Libraries: scikit-learn, TensorFlow, Keras, PyTorch, Langchain
• Data Analysis and Visualization: Pandas, NumPy, Matplotlib, Seaborn, Apache Spark, Opensearch, Elasticsearch (ELK), Splunk
• CAE Pre and Post processing : ANSA, Animator4
• CAE Solvers - LS-DYNA, Abaqus
• Cloud Platforms: AWS
• Version Control: Git, GitHub
• Containerization: Docker, Kubernetes
• Agile Development: Scrum, Kanban

Projects

• Kinematic evaluation and Injury prediction for 50th percentile Female CHBM | University of Stuttgart | Jul 2022

  • Repositioning of Viva Open HBM and simulation of the model with a Frontal Impact Load case.
  • Injury assessment of the HBM.

• Audio based predictive maintenace using Machine Learning | University of Stuttgart | Sep 2021

  • Developed a classifier using ANNs and CNNs to classify defective and good components of a machinery when audio recordings of the components under operation are provided.
  • Applied the model to perform predictive maintenance.

• Thermo-hyperelastic coupled problem for Large deformations | University of Stuttgart | Jul 2021

  • Numerical formulation of non-linear hyperelastic materials considering large deformations.
  • Object oriented implementation of the model in an in-house python FE-code called ez-FEM (Newton-Raphson solver)

Languages

• German (Very good - B2)
• English (Fluent - C1)

Interests

• Generative AI
• Integration of AI with CAE
• Finite Element Analysis - Automobile Crash, Pedestrian protection and Occupant Safety
• Computational Human Body Modeling

---