Open to Forward Deployed Engineer roles
Senior SRE / DevOps Engineer · AI & Platform
Hi, I’m Deepak Kumar’s AI assistant. Here’s Deepak in one minute.
Deepak is a Senior SRE and DevOps engineer at Nagarro in Gurugram, India, with nearly four years of keeping cloud platforms reliable, secure and cost-efficient.
He runs AWS and Kubernetes platforms for global R&D teams across five regions: 99.9% uptime for more than 5,000 users, deployments 40% faster and cloud costs down 15%, built on Terraform, AWS CDK and solid CI/CD.
When production breaks, he leads the response. He has handled more than 50 P1 and P2 incidents and cut recovery time by 20% with better observability.
He also builds AI tools for engineers, from MCP integrations to a DevOps automation platform on AWS Bedrock.
He’s open to Forward Deployed Engineer, Senior SRE and platform roles. Scroll down for the details, or ask me anything.
I embed with engineering teams to ship AI tools they actually adopt: MCP integrations, Bedrock agents and RAG for operations, running on AWS and Kubernetes platforms I keep reliable, secure and cost-efficient.
Now building: AI agents on AWS Bedrock, MCP tool integrations, RAG for operations, reliable Kubernetes platforms
AI engineering
Not demos: assistants, agents and detection systems built for real users, wired into the tools they already rely on.
Conversational AI platform connecting GitLab, Jira, Confluence and Miro through MCP, using AWS Bedrock for tool routing; reduced manual workflow tasks by 60%.
60% fewer manual workflow tasks
One assistant that gives engineers access to Black Duck, Snyk, GitLab and Miro, plus a Nexus MCP integration for AI-assisted engineering workflows.
Built for global R&D teams
Real-time detection and automatic revocation of compromised Kubernetes service-account tokens using multi-modal behavioral analysis; 95%+ detection accuracy with zero false positives in testing and under 3 seconds from detection to mitigation.
95%+ detection · <3 s to revoke
Retrieval-augmented support platform indexing 1,000+ knowledge articles for 500+ concurrent users; 85%+ query accuracy and 40% fewer support tickets, with SHA-256 audit logging, cost analytics and data sanitization aligned to SOC 2 and GDPR.
85%+ accuracy · 40% fewer tickets
An AWS Lambda and S3 chatbot deployed for a user base of 5,000+.
Serving 5,000+ users
The chat on this page: streams answers from my Bedrock gateway (Docker on EC2 behind CloudFront) with a Gemini fallback, strict input validation and Redis rate limits, and passes recruiters’ messages to me.
Try it now
Designs LLM tools around token usage and cost, and ran an internal AI community session on practical adoption and engineering productivity.
Connects models to GitLab, Jira, Confluence, Miro, Nexus, Snyk and Black Duck through MCP, so the assistant works where people already work.
Secret and token lifecycle management, least-privilege IAM and RBAC, and automated revocation of compromised credentials.
Forward Deployed Engineering
I haven’t held the FDE title yet, but the work is what I already do: embed with users, find the real problem, and ship working software into their environment.
Embedded with global R&D teams across the Americas, EMEIA, China, Finland and the UK, owning the platforms and tools they depend on
Builds working tools directly for users: internal R&D bot (Black Duck, Snyk, GitLab, Miro), Nexus MCP integration for AI-assisted workflows, Lambda/S3 chatbot for 5,000+ users, self-service Miro provisioning, 100+ Atlassian Forge macros
Hands-on with customer-facing delivery: migrations with zero data loss (700+ Atlassian users, GitLab instances) and zero-downtime AWS cutovers
Presents to senior leadership and global stakeholders; runs knowledge-sharing and an internal AI adoption session; mentors engineers (topmate.io, Preplaced)
Builds AI systems end to end: enterprise DevOps AI automation platform (MCP + AWS Bedrock), RAG support platform, AI-driven Kubernetes token-theft detection
For engineering leaders
Embeds with users to build tools they actually use: MCP integrations, Bedrock agents and RAG for operations.
60% fewer manual workflow tasks · FDE-ready
SLIs, SLOs and alerting that catch problems early, plus calm incident leadership when they don’t.
99.9% uptime · 95% SLA compliance · 99.95% on EKS
Reusable Terraform and AWS CDK modules and CI/CD pipelines that make shipping routine.
40% faster deploys · 15% lower spend · 25% fewer rollbacks
DevSecOps gates, least-privilege IAM and RBAC, and automated response to compromised credentials.
SonarQube · Snyk · Black Duck · token revocation in <3s
Senior SRE / DevOps Engineer who keeps cloud platforms reliable, secure and cheap at scale: 99.9% uptime for 5,000+ users across 5 global regions, 15% cloud cost cut, 40% faster deployments. Over 3+ years at Nagarro he has owned reliability and platform engineering for AWS + Kubernetes (EKS) environments supporting mission-critical R&D tools across the Americas, EMEIA, China, Finland and the UK, with 95% SLA compliance. Strong ownership under pressure with a bias for automation: if a task is manual and repeated, he turns it into a pipeline, a module or a self-service tool. Enjoys mentoring engineers, running knowledge-sharing sessions and presenting to senior leadership and global stakeholders.
Experience
October 2026 – present · Gurugram
July 2024 – September 2026 · Noida
March 2024 – June 2024 · Gurugram
March 2023 – March 2024 · Gurugram
More projects
2024
ELK observability platform with Logstash parsing and error classification, plus Kibana dashboards and alerts for error spikes, API failures, downtime, disk usage and indexing delays.
2025 – 2026
Production-style NestJS API behind his website with validation, rate limiting, health checks, structured logging and Swagger.
2024
Content-publishing platform with social features; PostgreSQL indexing and query tuning cut response time 60% and Redis caching cut database load 45%.
2024
Full-stack gym platform with JWT auth and role-based access, live check-ins over Socket.io, AWS SES email, and Docker Compose with Nginx.
Skills
Credentials
B.Tech in Computer Science
Aliah University, Kolkata, West Bengal, India
August 2019 – May 2023
Hindi (native or bilingual) · English (full professional)
Open to full-time roles and selective freelance/contract work; remote worldwide; open to relocation.