Looking for a job

Kumar Chodavarapu

@kumarchodavarapu

Message

I’m an MLOps engineer and ML engineer building GPU-optimized, end-to-end production ML platforms.

Canada

Message

What I'm looking for

I’m looking to lead end-to-end MLOps and LLM production—designing GPU-efficient Kubernetes platforms, CI/CD, monitoring, and drift-driven retraining—where engineering culture values reliability, security, and measurable business impact.

I’m an MLOps engineer with 12+ years of progressive experience, starting in core IT systems engineering and evolving through DevOps, ML engineering, and now MLOps. I intentionally carry the Linux, automation, and infrastructure foundation into every ML system I build.

Since 2019, I’ve applied that depth to real AI/ML workloads—building telecom network ML models at Ericsson using real client datasets, then leading full MLOps platform engineering at Qubittron. I’m an expert in end-to-end ML lifecycle management, GPU-optimised infrastructure, LLM/GenAI serving, and production model monitoring.

At Qubittron, I designed and led a central ML platform used by 10+ data science teams, taking projects from experiment tracking to live monitoring. I built GPU-optimised Kubernetes clusters on AWS EKS, cutting GPU idle time by 40%, and automated Continuous Training with Kubeflow and MLflow—reducing model-to-production time from 3 weeks to under 6 hours.

I also engineered high-throughput real-time serving with KServe and Seldon Core (10M+ predictions/day at p99 latency under 80ms) and delivered an LLM inference layer with vLLM and HuggingFace TGI. I implemented RAG and vector search pipelines (FAISS/Pinecone, LangChain), feature stores (Feast), and drift detection with Evidently AI and Kafka—triggering retraining and reducing degradation incidents by 75%, while embedding security and governance practices like RBAC, Vault-based secrets management, and SOC2-aligned audit logging.

Experience

Work history, roles, and key accomplishments

Current

MLOps Engineer

Current

Qubittron Consulting Inc

Apr 2022 - Present (4 years)

Designed and led a central MLOps platform for 10+ data science teams, reducing model-to-production time from 3 weeks to under 6 hours. Built GPU-optimized AWS EKS clusters and real-time LLM serving (10M+ predictions/day) while cutting GPU idle time by 40% and reducing model degradation incidents by 75%.

Kubernetes Kubeflow MLFlow KServe VLLM Evidently AI Kafka Faiss

ML & DevOps Engineer

Ericsson

Apr 2019 - Feb 2022 (2 years 10 months)

Managed telecom cloud platforms and built ML models on real client network datasets across 50+ production clusters. Trained anomaly and fault classification models achieving 88%+ precision and built an LSTM-based predictive maintenance model forecasting failures up to 48 hours in advance.

Kubernetes OpenStack Terraform CloudFormation Prometheus Grafana scikit learn XGBoost

DevOps Engineer

Videri North

Jul 2017 - Jan 2019 (1 year 6 months)

Automated infrastructure provisioning with Terraform and CloudFormation, reducing deployment time by 40% and configuration errors by 25%. Built CI/CD pipelines with Jenkins and Git, deployed containerized workloads on AWS ECS with autoscaling, and improved peak traffic handling by 50%.

Terraform CloudFormation Jenkins CI CD Pipelines AWS ECS AutoScaling Kafka ELK Stack Prometheus Akamai CDN

Collaborative Services Analyst

Invesco

Jun 2010 - Jan 2017 (6 years 7 months)

Provided core Linux and application/server administration, including RedHat Linux management and troubleshooting of BEA WebLogic on Solaris, RHEL, and Windows. Led migrations (AIX to RedHat/Solaris), administered SAN/NAS storage, and wrote Python/Shell automation to monitor logs, disk space, and services.

RedHat VMWare Shell Scripting Python NIS Powerbroker

Systems Engineer

NESS Technologies

Jan 2005 - Mar 2010 (5 years 2 months)

Built AWS infrastructure automation using CloudFormation (VPC/subnets/NAT) and implemented security via IAM and Security Groups. Delivered continuous delivery with Bamboo/Bitbucket/Maven, integrated Chef for provisioning, and deployed monitoring with Splunk and AppDynamics for microservices.

AWS CloudFormation AWS Security Groups Docker AWS Lambda AWS S3 Jenkins Splunk AppDynamics