Architecting
Resilient Systems.
Orchestrating Cloud-Native & Agentic AI Infrastructure.
I don't just configure tools; I engineer systems. Bridging the gap between application logic and cloud-native infrastructure through first-principles Systems Thinking. Currently pioneering Agentic Workflows on Amazon Bedrock.
Building the Future of Cloud & AI
Platforms
"If a system can't tell you what it's doing, it's already broken—you just haven't been paged yet."
I'm Vishal, an SRE-focused engineer working on the infrastructure layer that sits between cloud platforms and the teams who build on top of them. I focus on DORA metrics and building systems that improve reliability and developer experience.
More recently, I’ve been exploring the intersection of Infrastructure and AI, experimenting with agentic workflows on Amazon Bedrock using the Model Context Protocol (MCP), and building assistive systems where LLM-based tools can help analyze failures and surface actionable insights.
I strongly believe that reliability is a feature. My focus is on building platforms that are easy to operate, observable, and practical to scale - while improving lead time for changes and time to detect issues through continuous iteration and real-world testing.
Cloud Platform Engineering
Architecting multi-AZ EKS clusters with modular Terraform, declarative ArgoCD GitOps, and Istio service mesh. Infrastructure that provisions itself, heals itself, and documents its own state.
SRE & Reliability
Treating uptime as a first-class engineering requirement. Full observability stacks, error budget tracking, runbook-driven incident response, and DevSecOps pipelines with zero-CVE deployment gates.
Agentic AI Infrastructure
Building the next layer of ops: LLM-powered agents on Amazon Bedrock that analyze build failures, correlate distributed traces, and orchestrate infrastructure changes via the Model Context Protocol.
Location & Languages
Based in Pune, India. English · Hindi · Marathi. Open to remote collaboration on cloud, DevOps, and AI infrastructure projects globally.
Skills & Technologies
Deep expertise across the full cloud-native stack - from bare-metal infrastructure to AI-driven agentic systems.
CORE Engineering Foundations
Systems-first engineering grounded in OS internals, DBMS architecture, concurrency models, and protocol-level networking.
SRE & Reliability Engineering
Engineering resilient distributed systems through observability, failure analysis, tracing, and strict SLO governance.
Cloud-Native Platform Engineering
Architecting scalable, self-healing infrastructure platforms using Kubernetes, GitOps, Terraform, and distributed cloud-native patterns.
DevSecOps & Governance
Embedding security, policy enforcement, and compliance automation directly into cloud-native delivery pipelines.
AI & Agentic Infrastructure
Designing GenAI infrastructure and intelligent workflow orchestration through MCP-driven agentic systems.
Systems Thinking Approach (STA)
Applying first-principles engineering to reliability, scalability, distributed systems behavior, and infrastructure decision-making.
Infrastructure & Observability
A live view into the stack powering this platform. Built on cloud-native principles, automated by code, and monitored in real-time.
Systems Engineering & Operational Lifecycle
Architecting resilience across distributed systems. From scaling Kubernetes to deploying Agentic AI, every role is a chapter in building the 'Invisible Infrastructure'.
Verified Credentials
Industry-recognized certifications and ongoing learning paths in cloud-native architecture.
Education
Academic foundations that shaped my engineering mindset and technical depth.
Community Roles
Contributing to cloud-native ecosystems while architecting the next generation of reliable cloud and AI infrastructure.
Google Developer Group
Actively involved in Pune’s developer ecosystem through GDG, engaging in deep technical discussions around cloud-native systems, distributed architectures, and modern backend patterns.
AWS User Group Pune
An active contributor to the AWS community, engaging in real-world discussions around scalable cloud architecture, DevOps, and production-grade systems. Exploring Agentic AI workflows with Amazon Bedrock.
CNCF Community
Deeply aligned with the cloud-native ecosystem, working with Kubernetes, observability, and platform engineering. Implementing CNCF tools in real-world projects and GitOps workflows.
FinOps Community
Engaged in FinOps practices to bridge engineering with cost efficiency. Optimizing cloud spend across AWS and Kubernetes workloads for operational clarity.
Atlassian Community
Refining workflow automation and team collaboration. Leveraging Atlassian Rovo and AI to enhance knowledge discovery and DevOps lifecycle management.
Snowflake Community
Exploring modern data platforms and cloud data warehousing. Connecting backend systems and DevOps practices with evolving data engineering workflows.
“Learning, building, and growing through active community engagement in modern engineering.”
Community Engagement & Events
Participating in cloud-native meetups, workshops, and developer events while continuously learning in the DevOps, Cloud, and AI ecosystem.
TECH MEETUPS ATTENDED
GLOBAL/APAC CONFERENCES
CLOUD COMMUNITIES
ACTIVE SINCE
Live Mission Control Dashboard
A high-fidelity simulation of a production cloud environment. Track Golden Signals, monitor GitOps deployment streams, and manage error budgets in real-time. Syncs directly with the SRE Terminal.
IaC Blueprint
VIEW_ARCHITECTUREOperational Log
LIVE_FEEDSystem Design Philosophy
How I approach building systems that work reliably in the real world - shaped by hands-on experience across projects and production environments.
I don’t treat observability as an add-on. Before shipping anything, I make sure the system can explain what it’s doing. If it can’t, debugging becomes guesswork. Metrics, logs, and traces help me understand system behavior early - so when something breaks (and it will), I’m not starting from zero.
In distributed systems, failure is normal - not an edge case. I design systems with that in mind from the beginning. Retries, circuit breakers, and graceful degradation are part of the design - ensuring the system behaves predictably even under stress.
Code shows what was built, but not why. I focus on documenting the reasoning behind decisions - trade-offs, constraints, and alternatives. This helps both teammates and my future self understand systems without rethinking everything from scratch.
If I find myself doing something more than once, I automate it. Repetition leads to inconsistency and human error. From CI/CD pipelines to infrastructure provisioning, I focus on building systems that are repeatable, predictable, and require minimal manual intervention.
Reliability isn't a one-time setup; it's a daily practice. From small commits to large architectural shifts, every decision is measured against its impact on the 99.99% availability goal.
Before abstracting complexity to Kubernetes or AWS, I analyze the underlying compute, OS thread management, and raw network topology. Tools change yearly; TCP/IP, DBMS locks, and Linux cgroups do not. I apply a Systems Thinking Approach to every architectural decision.
“Applied thinking from real systems, not just theory.”
How I Debug Production Systems
A structured, hypothesis-driven approach to debugging distributed systems - focusing on signals, narrowing scope, and reaching root cause without guesswork.
Identify the Symptom
I start from the edge - typically ingress or API gateway logs (Nginx/Envoy) - to understand what users are actually experiencing. I look for clear signals like HTTP 5xx errors, latency spikes, or dropped connections, and correlate timestamps to establish when the issue started.
Scope the Blast Radius
Before diving deeper, I determine whether the issue is isolated or systemic. Using metrics from Prometheus/Grafana, I check patterns - CPU, memory, request rates, DB connections - to see if multiple services are affected or just a single pod/node.
Pinpoint the Bottleneck
Once scoped, I follow the request path using distributed tracing (Jaeger/Datadog). This helps identify the exact service or span causing the issue - whether it’s a slow downstream dependency, a failing database query, or unexpected latency between services.
“Debugging with signals, not assumptions.”
DevSecOps Enterprise Pipeline
An end-to-end cloud-native platform for a 3-tier MERN application on Vercel - code commit to production with security gates at every stage. Click any stage to inspect it.
Production Architectures & Lab Deployments
Real-world DevOps, cloud infrastructure, and AI platform projects - automation, scalable architecture, and modern cloud-native engineering.
SEV-2 Post-Mortem: Kernel Cgroup Pressure
🧩 **The Problem:** Cascading OOMKills in the Retail API cluster triggered ~45% error rate during a traffic spike. Analysis revealed Linux kernel cgroup memory limit saturation and aggressive page-faulting under high concurrency. ⚙️ **The Architecture:** AWS EKS, Karpenter, Kyverno, Prometheus. 🧠 **Decision Log:** Isolated memory-leaking nodes and analyzed kernel-level resource constraints. Tuned Karpenter provisioning buffers and introduced Kyverno admission policies to enforce strict cgroup memory limits and QoS classes. 📊 **Production Impact:** Recovered service stability in ~20 minutes and established deterministic guardrails to prevent kernel-level resource misconfigurations.
High-Availability Retail Platform
🧩 **The Problem:** Distributed systems complexity and L4/L7 routing bottlenecks caused intermittent data loss and high P99 latency during peak traffic. ⚙️ **The Architecture:** AWS EKS, Istio Service Mesh, RabbitMQ, Redis, PostgreSQL (Multi-AZ). 🧠 **Decision Log:** Architected an asynchronous event-driven system to decouple core domains. Implemented Istio for advanced traffic management (Circuit Breaking/Retries) and optimized PostgreSQL connection multiplexing via server-side pooling. 📊 **Production Impact:** Reduced provisioning time by 70% and achieved sub-10ms response times by solving distributed system race conditions and lock contention.
Enterprise DevSecOps CI/CD Platform
🧩 **The Problem:** Fragmented security validation and lack of observability led to inconsistent deployment quality and high MTTR. ⚙️ **The Architecture:** Jenkins, Terraform, SonarQube, Trivy, ArgoCD, ELK Stack, Prometheus, and Grafana. 🧠 **Decision Log:** Orchestrated multi-stage pipelines with rigorous security gates (SAST/CVE scanning) and centralized logging to intercept vulnerabilities pre-production. 📊 **Production Impact:** Slashed MTTR by 50% via immediate root-cause identification and managed FinOps tracking to monitor SLIs/SLOs and DORA metrics.
Articles & Engineering Insights
Sharing learnings on DevOps, cloud-native platforms, Kubernetes, AI infrastructure, and modern engineering practices.
What People Say
Feedback from peers, mentors, and community collaborators.
Working With Me
What you can expect when building critical infrastructure together - principles I ship with.
Engineering Reliable Systems For A Chaotic Internet.
Open to DevOps, cloud infrastructure, platform engineering, and collaboration opportunities.
"Calm infrastructure creates fast teams."
Building Bridges in Pune's Tech Ecosystem
I believe the best infrastructure is invisible—the silent force that empowers developers and ensures a seamless experience. As an active participant in the Atlassian Community, I’m dedicated to sharing knowledge and refining the workflows that drive our industry forward.
