Lohit Kolluri
Platform Engineer | SRE & Cloud Infrastructure
about
As a Platform Engineer | SRE & Cloud Infrastructure focused on reliability, automation, and infrastructure at scale. At Apollo Tyres R&D, I architected a high-performance backend managing 200+ concurrent simulation jobs, improving system throughput by 40% and reducing P95 latency by 35%.
My expertise spans cloud platforms (AWS, Azure), containerization (Docker, Kubernetes), and CI/CD pipelines. I build platforms and tools like KubeWise (an AI-driven Kubernetes SRE platform that cut MTTD by 60%), NLP2SQL (AI-powered database query engine), and enterprise simulation systems — each with measurable improvements in performance and reliability.
As a 3x Microsoft Certified, 5x Hackathon winner, and Microsoft Student Ambassador, I combine technical expertise with practical problem-solving. Currently pursuing B.Tech in Computer Science at SRM University.
- AWS
- Azure
- Kubernetes
- Docker
- Terraform
- CI/CD
- Linux
- Python
- FastAPI
- Django
- PostgreSQL
- MongoDB
- Redis
- RabbitMQ
- Celery
- React
- Next.js
work experience
- →Architected and implemented a high-performance backend system using Django and PostgreSQL, handling 200+ concurrent simulations with 40% improved efficiency
- →Created real-time dashboards using Chart.js for tracking project metrics across 300+ projects, improving operational visibility by 50%
- →Collaborated with cross-functional teams to optimize simulation workflows, resulting in 30% faster task allocation
- →Implemented automated testing and CI/CD pipelines, reducing deployment time by 45%
DjangoPostgreSQLREST APICI/CDDockerAzureChart.js
EduSkills Foundation AI Intern | Explored AWS Cloud & Machine Learning
- →Completed comprehensive AWS certification training, mastering deployment of ML models on AWS services (S3, ECS, Lambda)
- →Developed and deployed scalable cloud solutions, implementing best practices in security and architecture
- →Gained hands-on experience with cloud security, networking, and infrastructure management
AWSMachine LearningS3LambdaECSEC2
MathWorks Virtual Intern - MathWorks | AICTE NEAT Program
- →Completed advanced courses in AI tools and technologies provided by MathWorks
- →Developed proficiency in MATLAB for data analysis and model development
- →Applied machine learning concepts to solve real-world engineering problems
MatlabMachine LearningData Analysis
Edunet Foundation AI Intern | Machine Learning and Data Analysis
- →Developed a Mental Health Fitness Tracker achieving 98.50% accuracy using Python and scikit-learn
- →Implemented 12 regression algorithms for comprehensive mental fitness analysis
- →Enhanced model performance through advanced preprocessing and feature engineering
- →Optimized model parameters using ensemble methods for improved accuracy
PythonScikit-LearnMachine LearningData Analysis
projects
Reduced Mean Time To Detection (MTTD) by 60% and Mean Time To Resolution (MTTR) by 45% by creating a system to auto-detect, diagnose, and remediate cluster incidents using Prometheus and AI forecasting.
- FastAPI
- Prometheus
- Kubernetes
- Gemini
- MongoDB
Architected a multi-tenant backend infrastructure to support both PostgreSQL & SQLite, featuring zero-downtime schema switching for continuous availability.
- Python
- Azure OpenAI
- PostgreSQL
- Streamlit
Modernized the application life cycle by containerizing with Docker and automating deployments via a CI/CD pipeline, which resulted in consistent, zero-downtime releases.
- Django
- PostgreSQL
- Docker
- Azure
oss contributions
· Added explicit CUDA architectures for Pascal GPU support in llama.cpp build — enabling GTX 10-series and Tesla P40 compatibility
· Fixed the constraint enforcer to skip completed tasks — cleaner scheduling for swarm clusters
· Excluded completed tasks from constraint accounting — prevents false capacity blocks in SwarmKit
· Enabled metamorphic defaults in optimizer tests — randomized coverage at scale
· Shipped docs, validation, and test hardening across 4 PRs — production polish
· Documented compaction behavior that may temporarily exceed retention.size limits
· Parallelized 50+ util/app tests with t.Parallel() — faster CI feedback loops
· Fixed modal color bleeding with opaque border tokens — cleaner dark-theme UI
· Fixed UTF-8 charset serving for workspace text files — no more garbled output
