Ashish A Kamat
Most AI fails at the infra layer. Mine doesn’t. Staff AI Platform Engineer — Agentic IDPs, Kubernetes, Terraform | 60% less friction | Relocating across Germany/NL/EU| CET
Bengaluru
11+ years of work experience
Sep 20251 yrPresent
Bengaluru, Karnataka, India
• Architected enterprise AI platform infrastructure supporting internal generative AI agents and LLM applications using GCP, Kubernetes, Terraform, and Vertex AI.
• Built agent-based automation platform integrating GitHub, Jira, and ServiceNow to automate CI/CD troubleshooting and incident diagnostics.
• Implemented AI observability pipelines using Phoenix AI and Grafana to monitor model performance, token consumption, and request latency.
• Designed multi-region disaster recovery architecture across GCP services including Cloud Run, AlloyDB, and Redis.
Aug 20241 yr 2 mosSep 2025
Bengaluru, Karnataka, India
- Designed and Lead efforts of cost optimization framework for Kubernetes clusters running on EKS using a blend of open-source tools and in-house built tools. This helped the team and the overall company reduce costs on Kubernetes and overall AWS resources by up to 60%. (Currently working on the next generation of this using Agentic AI.)
- Designed and automated CI/CD pipelines for LLM-based services using GitLab CI, Terraform, and containerized workflows on AWS EKS.
- Owned GitLab management and created chatbot-driven automation for auto-approving changes of repository ownership, merge request approvals, runner onboarding, etc.
- Collaborated closely on a daily basis with Data Engineers and Data Scientists to enable seamless deployment of GenAI applications, including fine-tuned LLMs and RAG-based solutions.
- Lead efforts of Implementation of AWS CloudWatch synthetic canary dashboards and end-to-end automation processes such as Infrastructure as Code using Terraform, GitLab pipelines, and parameterization for flexibility across services—enhancing system and application performance, including monitoring stack (Prometheus + Grafana), by 60%.
Jan 20231 yr 8 mosAug 2024
Hybrid
- Implemented Git-ops for Kubernetes-based machine learning applications, reducing deployment errors by 50%.
- created GitOps based deployment using ArgoCD for inhouse cloud-native Kubernetes platforms, achieving 40% faster time-to-insight with optimized data pipelines and real-time processing. Created custom Helm charts and pipelines, achieving 25% less downtime with containerization and collaboration with DevOps engineers, improving user experience.
Oct 20211 yr 3 mosDec 2022
Remote
Part of the Centralised SRE team who is responsible for the development, enhancement, and support for Walmart's own proprietary CI-CD tools such as looper, concord (https://concord.walmartlabs.com), oneops, and WCNP (Walmart cloud-native platform)
- designed, architect, implemented, and onboarded multiple teams/ applications to follow the GitOps way and rolled our ArgoCD for Kubernetes-based deployment
- designed, architect, and onboarded a couple of teams from Looper (Walmart's own CI tool) to Gitlab by creating end to end to pipeline creation, project administration, and Integration with other Walmart's CI/CD tools such as Concord, Github
- designed, architect, implemented, and migrated monitoring dashboards of various Jenkins/Looper dashboards to the single centralized dashboard with the creation of custom and standard metrics which helps an on-call person to troubleshoot issues easily
- designed, architect, and implemented a Single source of the dashboard to monitor Infrastructure getting used behind all CI/CD tools and removing the dependency of Bastion host to run ad-hoc Ansible tasks by giving UI trigger feature.
- researched and implemented Configuration as Code feature for Looper by which UI tasks that are getting performed by a part of CRQ can be eliminated with the use of filters and CLI commands
- Helping various product teams to migrate their monolithic applications to microservice-based architecture by the creation of Helm charts, and customized pipelines suitable for deployment on WCNP, a multi-cloud cloud-native Kubernetes platform
- as part of the cross-team member, written various custom CRDs and controllers to support cloud-native deployments on WCNP
Apr 20217 mosOct 2021
Mumbai, Maharashtra, India
As a senior member of the centralized DevOps team, leveraged expertise in AWS DevOps, Kubernetes, Terraform, GitLab, serverless frameworks, and observability tools to optimize serverless deployment times, resulting in a 20% improvement in deployment speed and 50% reduction in downtime, while ensuring 99.99% uptime and real-time visibility into application performance.
Jun 20191 yr 11 mosApr 2021
Mumbai Metropolitan Region
As Senior Member of a team, implemented a DevOps platform and successfully onboarded 70+ microservices onto Kubernetes, reducing production incidents by 40%, improving mean time to resolution by 50%, achieving a 50% reduction in deployment time, ensuring 99.99% uptime and real-time monitoring of application performance, and optimizing data processing by 40% through the implementation of machine learning pipelines.
Apr 20181 y 4 moAug 2019
Mumbai
Mar 201711 moFeb 2018
Thane
Sep 20142 ySep 2016
Pune
Ashish’s AI Stack
Education
Master's degree in Science, Data Science
Liverpool John Moores University
20241 yr2025
Postgraduate Degree, Data scientist
International Institute of Information Technology Bangalore
20221 yr2023
Join Ashish and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal Boss