Vihar Devalla
vihar@five-nines:~$ cat about.md
about
I build systems that stay up.
I own the reliability of financial infrastructure that processes billions in daily transactions. My job is to make the platform invisible — the kind of reliable where nobody notices it.
I came up through cybersecurity and systems research before moving into SRE. That background means I approach production systems as an adversary would: asking where they break before they do.
location: Bengaluru, India
certs: CKA + CKS (Linux Foundation)
focus: Kubernetes, Observability, Platform Engineering
open_to: Senior SRE / Staff SRE / Platform Eng roles
Error Budgets Over Uptime Theater
Reliability is not about achieving 100% uptime — it is about defining, measuring, and defending a promise. I work in error budgets, not vanity metrics.
On-call Is a Feedback Loop
Every page is a signal. Incidents are not failures to be ashamed of — they are the most honest data about system behavior. I treat them as learning opportunities.
Every Runbook Is Future Automation
If I am doing the same thing twice, I am doing it wrong the second time. Toil reduction is not laziness — it is engineering discipline.
vihar@career.log:~$ kubectl get jobs --sort-by=.status.startTime
experience
SRE II
Platform Site Reliability Engineering
@ Arcesium
Bengaluru, India
▸ key initiatives
Disaster recovery across regional cloud failures and ransomware scenarios — coordinating cross-functional teams with a clear targeted RPO.
Designed and built a firm-wide scoring system enabling stakeholders to assess observability posture and application risk across the platform.
Optimized platform observability and built automation tools that accelerate root cause identification and faster incident resolution.
— previous roles —
vihar@tool-shed:~$ kubectl describe skills --output=wide
skills
vihar@brag-file:~$ cat achievements.yaml
achievements
certifications
speaking
awards
$ kubectl get observability-stack --all-namespaces
observability stack
Metrics, traces, and dashboards for platform services. Alerting on golden signals: latency, traffic, errors, saturation.
Elasticsearch + Logstash + Kibana for structured log ingestion, search, and anomaly detection across platform services.
Custom dashboards for SLO/SLI tracking, capacity planning, and incident correlation. Used for release readiness reviews.
Built a platform integration for Signoz at the Arcesium Season of Code 2024 — bringing metrics, logs, traces, and events into one place. Won the hackathon.
reliability focus areas
$ ls incident_patterns/ | sort -u
incident patterns
// Real failure categories encountered. No fabricated IDs or invented numbers.
vihar@side-effects:~$ ls -la projects/ --show-hidden-impact
projects
Reliability Risk Audit Scoring System
Template-based framework in Go that pulls risk auditing metrics from various sources across the firm. Includes a web dashboard to visualize data. Used firm-wide to approve and gate new releases, and to help application teams adopt reliability as part of the SDLC cycle.
Kubernetes Debugging Tool for Slack
Slack bot using the Slack-Bolt Framework that retrieves Kubernetes resource health and surfaces logs, metrics, and network paths directly via Slack. Enables faster diagnostics without kubectl access, improving developer productivity and incident response speed.
Multi-Factor Authentication Portal (IAM)
Secure, scalable MFA self-service portal using Ping Identity APIs and clustered deployments to manage enterprise MFA device workflows. Built for high-availability with zero trust principles.
Unified Observability Platform — Signoz
Won the Arcesium Season of Code Hackathon 2024 by building a platform integration for Signoz — bringing metrics, logs, traces, and events into one singular place. Demonstrated the value of unified observability vs. fragmented tooling.
vihar@published:~$ ls papers/ | wc -l && cat each
publications
Temples Restoration using Gated Convolution and Contextual Attention in GANs
Novel deep learning approach to digitally restore broken and eroded heritage temples and sculptures in India using Gated Convolution and Contextual Attention layers in GANs.
mURLi: A Tool for Detection of Malicious URLs and Injection Attacks
Tool for detection of malicious URLs, SQL Injection, and NoSQL Injection attacks using ML and Deep Learning models.
Analytical Comparison of Models for Raga Identification in Carnatic Classical Audio
Research on Indian Classical Music classification based on Raga using Machine Learning and Deep Learning models.
vihar@not-on-call-rn:~$ ./connect.sh
contact
vihar@not-on-call-rn:~$ send_message --to vihar
// Discussing SRE, reliability, or open to new opportunities
Open to Senior SRE, Staff SRE, and Platform Engineering roles. Also happy to talk shop about Kubernetes, observability, or incident response.
built with Next.js · designed for reliability