Divesh Kumar

Experience

• 

  IEMLabs

  Mar 2020 - Jun 2020

  Machine Learning Intern
• 

  MedTourEasy

  Oct 2020 - Nov 2020

  Data Engineer
• 

  The Sparks Foundation

  Oct 2020 - Nov 2020

  Data Science Intern
• 

  Cognizant

  Feb 2022 - Apr 2022

  Trainee

  started as trainee

• 

  Razor Edge Analytics

  Apr 2022 - Mar 2024

  Full-stack Developer

  As a Full Stack Developer, I have been instrumental in various aspects of our software development process.- Proficiently worked with Python Flask, facilitated machine learning integration, and skillfully implemented React.js andRedux for dynamic front-end solutions.- Developed over 100 APIs, all of which have been successfully deployed in production environments.- Played an active role in front-end development, which involved designing and developing user interfaces (UI/UX).- Actively contributed to the testing and deployment of our products to staging and Production environments. Show less

• 

  Innovatics

  Apr 2024 - now

  Solution Engineer

  - Private LLM Service: - Developed a private Large Language Model (LLM) service for querying personal data through achat widget. Led both backend and UI development.- Implemented Serverless ML Training: Leveraged ECS Fargate for running containerized ML models, eliminating theneed for server management. Ensured seamless dependency management and environment configuration within Dockerimages for efficient deployment.- Scalable and Reliable ECS Tasks: Configured ECS Service Auto Scaling to dynamically adjust task count based on CPUutilization, and used AWS Application Load Balancer for even traffic distribution. Set up health checks and proper resourcelimits for reliability.- SDK for ML Job Execution: Developed an SDK with Python to communicate securely with API endpoints, employing APIkey-based authentication and role-based access control for authorization.- High Availability and Secure EC2 Instances: Deployed EC2 instances in an Auto Scaling group across multipleAvailability Zones, applied security groups, enabled encryption, and regularly updated instances to ensure high availabilityand security.- Modular Flask Application for Predictions: Structured the Flask app using blueprints for modularity, managed modeldependencies, and optimized performance with caching and load balancing, deploying the app in a Kubernetes cluster forscalability.- Dynamic Deployment: Implemented dynamic deployment solutions using Docker, serverless computing, monitoring tools,storage solutions, and dynamic security configurations. Show less