Grid Powering the dark
  1. WebGPU adapter
  2. Allocating renderer
  3. Compiling shaders
  4. Laying out grid
  5. Calibrating spotlight
  6. Frame 0
Probing WebGPU… 0%

Senior Site Reliability Engineer

Making production
boring.

Distributed systems, AWS, DevOps, observability, and production engineering — 15+ years of keeping high-traffic platforms quiet, predictable, and fast. Currently shipping WebGPU experiences on the side.

AWSKubernetesOpenTelemetrySplunkTerraformGoPython
Years in production
0+
Availability SLO
0%
MTTR reduction
0%

01 · Signal

A signal from production.

I'm Sudeep — a Senior Site Reliability Engineer with 15+ years across financial and educational enterprise systems. The work is unglamorous on purpose: SLIs and SLOs, error budgets, blameless postmortems, OpenTelemetry traces, and Splunk pivots that turn a 5am page into a 5-minute root cause.

SRE Leadership

Expert in SRE best practices including SLIs, SLOs, error budgets, incident response, root cause analysis, production hardening, and automation at scale, with strong emphasis on OpenTelemetry-based monitoring, Splunk-driven operational analytics, and AI-augmented engineering workflows.

Impact

Proven record of improving availability by 14%, reducing operational toil through Infrastructure as Code (Terraform), and strengthening system resilience using observability-driven engineering and Chaos experimentation.

Core competencies

Cloud & Architecture

AWS Certified Solutions ArchitectEC2ECSLambdaS3RDSCloudWatchHigh AvailabilityDistributed Systems

Testing & Performance

LoadRunnerQTPCapacity PlanningScalability TestingJFRVisualVM

Observability

OpenTelemetry (OTEL)PrometheusGrafanaELK StackSplunkDynatraceNew RelicCloudWatch

Automation

Terraform AssociateCloudFormationAnsibleJenkinsGitOpsDockerPythonBash

SRE Practices

SLIsSLOsError BudgetsIncident ResponseRCABlameless Postmortems

AI-Augmented Engineering

AI-assisted DevelopmentPrompt-driven PrototypingAI-assisted Incident TriageWorkflow Automation with AIDocumentation Acceleration

02 · Track record

Building resilient systems.

Key clients

* Engagements via USM Business Systems

  1. Aug 2018 – Present

    Senior Performance / DevOps Engineer

    USM Business Systems – Fannie Mae

    • Architected and provisioned AWS infrastructure using Terraform and CloudFormation, increasing platform availability by 14%.
    • Established and operationalized SLIs, SLOs, and error budgets, improving service reliability by 18%.
    • Implemented OpenTelemetry instrumentation for distributed tracing and metrics across microservices, expanding observability coverage and accelerating incident triage.
    • Led Splunk-based telemetry and log analytics to detect anomalies, correlate incidents, and drive measurable reliability improvements.
    • Applied AI-assisted engineering workflows to streamline troubleshooting playbooks, automate operational documentation, and accelerate reliability delivery.
    • Leveraged AI coding assistants (GitHub Copilot, Cursor) to design, build, and ship production-grade web applications — demonstrating end-to-end AI-driven development from architecture through deployment.
  2. May 2015 – July 2018

    Senior Performance Test / Systems Automation Engineer

    USM Business Systems – College Board

    • Optimized distributed scoring platforms handling high-volume ETL workloads for nationwide assessments.
    • Diagnosed JVM memory leaks, thread contention, and CPU bottlenecks using JFR and VisualVM.
    • Improved system scalability through JVM tuning, infrastructure optimization, and workload modeling.
    • Designed end-to-end reliability and performance strategies for data-intensive systems.
  3. Aug 2012 – May 2015

    Senior Performance Test Engineer

    Remote Tiger Inc / USM – Fannie Mae

    • Developed enterprise reliability and performance strategies for mission-critical web applications.
    • Built automated test frameworks and workload simulations using LoadRunner and QTP.
    • Delivered executive dashboards with reliability KPIs and capacity insights.
  4. Pre-2012

    Early Engineering Journey

    Detroit Technologies → Enterprise Financial Systems

    Started with foundational systems and performance engineering work, then scaled into large enterprise production environments and SRE leadership.

03 · Foundation

Education & certifications.

Education

Master of Science in Software Engineering · Bachelor of Technology in Computer Science

Certifications

AWS Certified Solutions Architect – Associate · HashiCorp Certified: Terraform Associate

Experience

15+ years across Performance Engineering, SRE, and DevOps.

Education

Master of Science in Software Engineering.

Certifications

AWS Solutions Architect · Terraform Associate

Key Clients

Fannie Mae · College Board · Dow Chemicals

04 · Showcase

Visual WebGPU work.

Generative pieces where the engineering is the design. No off-the-shelf shaders, no canned scenes.

Nebula Forge

Live web app · WebGPU · Generative graphics

2026 – Present

A single-page WebGPU experience: a real-time market wire streamed through an audio-reactive cosmic particle field, simulated entirely on the GPU. Mono-first, no off-the-shelf shaders.

WebGPU WGSL compute Audio-reactive Real-time market data

05 · Open source

Things I built.

CredVigil

Open-source · Go · DevSecOps

2025 – Present

An open-source credential secrets scanner written in Go. Scans codebases, config files, git history, and live file changes for exposed credentials across 75+ platforms using triple-signal detection.

369 detection rulesShannon entropyBPE scoringZero-trust pipelineGit history scanningGo
  • Triple-signal confidence scoring — regex pattern matching (369 rules across 75+ categories), Shannon entropy analysis, and BPE token efficiency scoring combined into a 0–100% confidence score per finding.
  • Five-stage zero-trust post-processing pipeline: hash → redact → enrich → fingerprint → sanitize. Raw secrets never written to disk or output.
  • Modular architecture with five independently tested components: detection engine, secure pipeline, git integration, file system watcher, and internal event bus.
  • 230+ enumerated SecretType constants covering cloud providers, AI/ML platforms, payment processors, CI/CD tools, and performance testing platforms.
  • 14 end-to-end tests; full suite passes with Go race detector enabled.

Nebula Forge

Live web app · WebGPU · Generative graphics

2026 – Present

A single-page WebGPU experience: a real-time market wire (news, FX, crypto, M&A) streamed through an audio-reactive cosmic particle field. Simulation runs entirely on the GPU.

WebGPUWGSL computeAudio-reactivePost-processing graphReal-time market dataMono-first
  • GPU-resident particle simulation — every particle's position, velocity, and color is computed in a custom compute kernel each frame.
  • Multi-pass post-processing graph: bloom, chromatic aberration, beat-synced glitch passes composited on the GPU.
  • Real-time market wire drives the scene — the data feed is the visualization's input signal, not decoration.
  • Strict no-off-the-shelf-shaders policy — the entire visual identity lives in a couple hundred lines of hand-written shader code.

06 · Contact

Let's talk.

Open to Senior SRE, Platform Engineering, and Production Engineering roles.

07 · Voices

What others say.

Recommendations from clients, peers, and managers — sourced from LinkedIn.

Built mono-first @svemulapati · WebGPU · TSL