Sherwin Rajkumar

Science + Computing AG

• RAG: Building Retrieval Augmented Generation (PDF Chatbot) systems - Based on open-source LLMs such as Llama and Mistral AI
• DevOps: Provisioning and management of Cloud HPC Environments - IaC - Terraform, Ansible, AWS/Azure
• Analytics, Observability and Alerting: Big-data analytics with Elasticsearch/Kibana, Prometheus/Grafana, Opensearch and Splunk in complex environments
• Time Series Forecasting: Development of Time Series Forecasting models for CAE license usage prognosis
• Training: Preparing and regularly conducting internal training sessions on the topic of Artificial Intelligence and Machine Learning

EDAG Engineering GmbH

• Data-driven thermal and hydraulic dimensioning of Disk Brakes with classical Machine Learning models like Random Forests and Gradient Boosting
• Feature Engineering / Exploratory Data Analysis (EDA)
• Predicting rough brake dimensions during the early development phase
• Accelerated the System-development phase by about 85% compared to a conventional approach

EDAG Engineering GmbH

• FE-Explicit Dynamic modelling of Body-in-white parts for Crash Simulations (Solver: LS-DYNA | Preprocessor: ANSA)
• Head-impact analysis of Bicycle helmets according to norms EN 1078 and EN 960
• Automation of frequently used ANSA routines with ANSA Python API

Institute für Konstruktionstechnik und Techniches Design, Uni Stuttgart

• Complete development of a Q-Learning model (Reinforcement Learning) to adjust the contour to Homogenize the pressure in a press-fit between a Shaft and a Hub joint according to DIN 7190
• Iterative optimization (based on Material properties) to achieve the same
• Creation of a dataset with optimal contours and prediction of contours for certain parameters with deep neural networks