Hey 👋, I'm

Tejus Vijayakumar

I am into training deep neural nets on large datasets

As a multidisciplinary Engineer with over 3+ years of experience at a leading bank, I led a team of 6 engineers to develop and deploy a machine learning solution that classifies customers into segments, improving lead data quality for the sales team. I also developed more than 12 payload-encrypted APIs handling approximately 1 million daily requests. My hands-on experience includes developing deep learning models from scratch, such as a 60 million parameter Transformer Model trained on a single Nvidia A100 GPU with 40GB memory for code summary generation, Deep Q-Network (DQN) for reinforcement learning, and Residual Network (ResNet) for image processing. Additionally, I have worked with image processing tools like OpenCV and stay updated on important recent advancements, including multimodal Large Language Models (LLMs) and Small Language Models (SLMs) that can be deployed at the edge for inference.

Graduated MSc Artificial Intelligence & Machine Learning from University of Limerick, Ireland.

see my resume

Work Experience & Education

Machine Learning Engineer @Tech Mahindra

July 2018 - May 2021

  • Developed and implemented multiple ML models to analyse lead data for customer segmentation.
  • Created detailed dashboards to monitor service requests and analyse lead data, including projections for performance improvements.
  • Collaborated extensively with data scientists, analysts, and infrastructure teams to develop and deploy robust, scalable machine learning solutions
  • Proficient in implementing CI/CD pipelines, leveraging DevOps practices, and managing MLOps workflows to streamline and automate development
  • Create build and release pipelines, containerized applications using Docker, and deployed on Kubernetes for streamlined development and testing
  • Architected and developed scalable, multi-component machine learning software platform, enhancing performance and enabling real-time data analysis
  • Developed and maintained real-time dashboards to provide customers with actionable insights and enhanced visibility into key metrics
  • Achieved 80% reduction in deployment time through automation in Python and shell scripting.
  • Developed 12+ encrypted APIs performing data transformations, orchestrations, and aggregation.
  • Experience in implementing symmetric and asymmetric encryption methods like AES-256, RSA etc.