Deependra Kumar's Portfolio

About Me

Introduction

Passionate Backend Developer with 1.5+ years of hands-on experience in designing, building, and deploying enterprise-grade scalable systems. Specialized in Python ecosystem with expertise in FastAPI, PostgreSQL, Redis, and cloud technologies including AWS (SageMaker, Lambda, Step Functions) and Google Cloud Platform (GKE).

Proven track record in microservices architecture, containerization with Docker and Kubernetes, and implementing comprehensive monitoring solutions. Known for taking end-to-end ownership of projects from conception to production deployment, with strong focus on performance optimization and system reliability.

Recognized for exceptional problem-solving abilities with HackerRank Gold badges in Python and Problem Solving, and recipient of multiple awards including the K Vasudevan Award and Best Team Award. Committed to continuous learning and contributing to high-impact projects in collaborative engineering environments.

Deependra Kumar

+91-9685468824

kdeep05.dg@gmail.com

www.linkedin.com/in/deependrakrdk

https://github.com/deependra0411

Professional Projects & Technical Leadership

Enterprise-Grade Backend Systems & Cloud Architecture

LLM-Powered Customer Support Chatbot - Sean

Nov 2024 - Present Production System

Live Demo

System Design & Scalable Backend Development

Built backend architecture from scratch using FastAPI, Redis, Redis Streams, and PostgreSQL, optimized to handle up to 2000 requests per minute (RPM)
Designed and implemented a modular, scalable backend structure to support complex business logic and performant database operations
Translated product designs and wireframes into production-ready code using FastAPI and PostgreSQL
Developed and deployed microservices with strong focus on scalability, modularity, and observability

Cloud-Native Deployment & Infrastructure (GCP)

Set up Google Kubernetes Engine (GKE) from scratch and managed end-to-end deployment of backend services using Docker and Kubernetes
Integrated Redis & Redis Streams for real-time message transformation and conversation state tracking across services
Enabled WebSocket communication at scale via Nginx Ingress Controller in Kubernetes, supporting real-time bidirectional flows

Auto-Scaling, reliability, Monitoring & Observability

Implemented Horizontal Pod Autoscaling (HPA) and custom Horizontal Pod Autoscaling (HPA) based on Prometheus metrics, dynamically scaling WebSocket services based on active connections
configure Pod Distruption Budget (PDB) to avoid down time during any disruption
Set up distributed tracing and logging using OpenTelemetry (OTEL) to monitor service health and debug across microservices
Established proactive alerting and anomaly detection using GCP Monitoring Alerts

CI/CD & DevOps

Set up CI/CD pipelines using GitHub Actions for automated building, and deployment

FastAPI Redis Streams PostgreSQL Kubernetes GKE Docker Prometheus OpenTelemetry WebSocket Nginx GitHub Actions

Serverless ML Inference Pipeline on AWS

May 2024 - Nov 2024 AWS Architecture

End-to-End Serverless Architecture

Architected and deployed an end-to-end ML inference pipeline on AWS for a retail use case, designed to process batch JSON inputs derived from CSV files uploaded to S3
Implemented a Lambda-triggered workflow with automated CSV to JSON conversion and timestamp-based file tracking

Automated Workflow & Concurrency Management

Designed workflow: Users upload CSV → Lambda converts to JSON → Second Lambda triggers 7 SageMaker endpoints serving distinct ONNX models → invokes step Function → final output
Managed concurrency using AWS Step Functions with max concurrency of 2 parallel branches, processing 500 records concurrently across all 7 endpoints
Implemented intelligent result aggregation with custom business rules and automated cleanup mechanisms

Fault Tolerance & Cost Optimization

Designed robust cleanup mechanism using Lambda functions to automatically destroy all 7 SageMaker endpoints after successful completion or failure scenarios
Ensured complete serverless, event-driven automation with strong emphasis on cost-efficiency, scalability, and fault-tolerance across all pipeline stages
Handled OOM errors, runtime exceptions, and timeout triggers with automated recovery

AWS Lambda SageMaker Step Functions Amazon S3 Event Scheduler ONNX Models Python Boto3 CloudWatch

Deependra Kumar

Software Developer

About Me

Deependra Kumar

Education

Bachelor of Engineering in Electrical Engineering

SHRI VAISHNAV INSTITUTE OF TECHNOLOGY AND SCIENCE (SVITS)

Achievements & Awards

Python Gold Badge

Problem Solving Gold Badge

SQL Silver Badge

Best Team Award

K Vasudevan Award

Certifications

Python (Basic)

Problem Solving (Basic)

Technical Skills

Backend Technologies

DevOps & Infrastructure

Cloud Platforms

Tools & OS

Soft Skills

Professional Projects & Technical Leadership

LLM-Powered Customer Support Chatbot - Sean

System Design & Scalable Backend Development

Cloud-Native Deployment & Infrastructure (GCP)

Auto-Scaling, reliability, Monitoring & Observability

CI/CD & DevOps

Serverless ML Inference Pipeline on AWS

End-to-End Serverless Architecture

Automated Workflow & Concurrency Management

Fault Tolerance & Cost Optimization

GitHub Statistics

150+

500+

Pull Shark x3

Public Repository Activity

Contact Me