Chethan Sanna Reddy

Worked on LLM and backend services for SDS automation with FastAPI, Temporal and PostgreSQL. Added provider-neutral routing that excludes providers storing or training on customer data, validated by a 132-case test matrix; built configurable LLM plugins, cost tracking and reliability fixes. Built Helacloud’s backend as its main developer, including SAP EHS integration, async workflows and API security. Expanded tests to 1,800+ and cut API response times by 75%. Also built an LLM benchmarking service.

Contributed to ML development for automated Safety Data Sheet (SDS) extraction, from PDF processing and BERT-based classification to production pipelines. Led a five-person annotation team, defining guidelines and quality checks. Used Doccano, Airflow and MLflow to connect labelling, training and evaluation, including an active-learning loop that cut manual labelling by up to 70% in pilots. Developed Python evaluation tools and supported model deployment with Docker and GitHub Actions.

• Detected a set of electric components and text from Single Line Diagrams • Implementation of algorithm to match the object with corresponding text • Focus: Microsoft Azure, LabelImg, Object detection, Text extraction

Tasks: • Researched on existing driver behavior models • Collaborated with other team to gather required CAN data • Classified (SVM, Decision Trees) road types with 97% validation accuracy • Clustered (K-means, Gaussian Mixture Model) different driver behaviours for each road type • Libraries: NumPy, Pandas, Matplotlib, Seaborn, Scikit-learn (sklearn)

Tasks: • Developed mathematical model for success probability • Implemented and compared Epidemic Routing algorithm’s results with mathematical model for different cache size • Concepts & Tools: Opportunistic Networks (OppNets), Matlab

Project: Towards a Smart Factory – Digitization for the Workshop Tasks: • Ran Scrum ceremonies and removed blockers to keep sprints on track • Enforced Scrum practices and aligned team, product, and stakeholders

Tasks: • Researched threat landscape and agents in connected cars • Proposed mitigation strategies and challenges in implementation and a framework for Business Resilience

• Manual testing (unit and integration), and a few configuration tasks with no issues in production environment • Created comprehensive test plans • Used SQL to retrieve, update and delete data from database • Mentored a junior colleague

• Manual testing and a few configuration tasks • Created test cases and test plans • Monitored production defects and fixed them • SQL queries to modify the database

• Performed comprehensive Regression testing • Monitored production defects and fixed them

Coursework: Convolutional / Recurrent Neural Networks (CNN/RNN), Generative Adversarial Networks (GAN), Natural Language Processing and Cloud Computing (AWS SageMaker, Lambda)

Coursework: Machine Learning and Statistics

Coursework: Data Structures and Object Oriented Programming in C++, Probability and Random Processes