abhiman@career:~$

Abhiman Kolte

Staff Software Engineer — Infrastructure, Reliability & Production Engineering

abhiman@career — zsh — 80×24
scroll to run pipeline ▾
// career_impact.json
0
change failure rate
0
on-call triage time
0
change systems gated
0
services onboarded
cat SUMMARY.md

Abhiman Kolte

Staff Software Engineer — Infrastructure, Reliability & Production Engineering · SF Bay Area, CA

Staff Software Engineer with 8+ years of software engineering experience, including 6+ years building reliability and developer infrastructure at LinkedIn and Amazon. Created LinkedIn's change safety platform from scratch — a Go and gRPC system now gating nine change agent systems, including the central deployment pipeline — cutting change failure rate by over 60%. Also owns SLI/SLO practice and capacity planning for the org.

.github/workflows/career.yml

CI/CD — Career Pipeline

Click run deploy on any stage to expand the job log.

#a1e4c9d
LinkedIn
Staff Software Engineer, Production Engineering (promoted from Senior Software Engineer)
Sunnyvale, CA
Jun 2022 — Present
ACTIVE · 99.99% uptime
$ deploy --target=linkedin --role="Staff Software Engineer, Production Engineering"
BUILD TEST DEPLOY
Seatbelt — change safety platform for production deploymentsGo · gRPC · Temporal
  • Created LinkedIn's change safety platform from scratch. Blocks production changes during system instability to prevent cascading failures — cut change failure rate by over 60% org-wide.
  • Engineered the policy engine as a gRPC service sustaining thousands of requests/sec under a strict sub-300ms budget, inline in the critical path of every gated change.
  • Designed a pluggable validator architecture so distributed teams author and own custom safety checks, moving safety policy to the engineers closest to each risk.
  • Integrated the platform across nine change agent systems, including the central deployment pipeline, then led the migration off the legacy Deployment Policy Engine to unify safety standards org-wide.
  • Built the rollout orchestrator on Temporal for durable, replay-safe execution across thousands of concurrent workflows, advancing one colo at a time with every transition re-checking the policy engine first.
  • Fed centralized alerting and OpenTelemetry signals into that decision, so a critical alert on the deploying service pauses its rollout mid-flight even when the broader environment is healthy — catching regressions the deployment itself introduced.
  • Eliminated race conditions during concurrent lock acquisition across independent deployment pipelines with a two-phase coordination protocol: a sync pre-check at workflow start plus async conflict detection at acquisition, inside a 5 second budget.
  • Shipped a fail-open architecture with independent kill switches and a shadow-enforce ramp, so platform outages degrade gracefully rather than blocking deployments.
Functional test automation platformPlaywright · API probing
  • Built the platform behind automated functional verification at LinkedIn, covering Playwright web E2E flows and an API prober. Onboarded 900+ services — over half the microservice fleet — to automated API testing, and 50+ business-critical services to frontend testing.
  • Made the suites callable at every stage of the change lifecycle: ad hoc before a pull request, as a pre-deploy gate, on a schedule for continuous monitoring, and inside canary — where a service's own tests run against the single host carrying the new version and trigger automatic rollback on failure.
Incident triage automation — agentic evidence collection & RCA
  • Delivered a production agentic triage service that cut on-call investigation from 15–20 minutes down to under 30 seconds per failure.
  • Built a multi-source evidence fusion layer correlating synthetic monitoring, L1 proxy logs (KQL), trace spans, and raw job logs into a grounded root-cause payload — keeping the endpoint deterministic and AI disabled for an orchestrating agent to call, with the retrieval strategy behind each decision recorded so every conclusion stays auditable.
SRE practice and tooling
  • Own SLI/SLO management, capacity planning, and service availability, running infrastructure at scale on Kubernetes and Terraform.
  • Wrote Confetti, a benchmarking library that profiles the performance impact of configuration changes inside CI/CD pipelines.
#9f7b2a1
Amazon Lab126
Software Development Engineer
Sunnyvale, CA
Nov 2020 — Jun 2022
COMPLETE · 99.99% uptime
$ deploy --target=amazon-lab126 --role="Software Development Engineer"
BUILD TEST DEPLOY
Monolith → microservices migration, core device platform
  • Owned architecture and delivery of a monolith to microservices migration for a core device platform, decomposing coupled subsystems into independently deployable services with defined API contracts, distributed tracing, and automated observability — holding 99.99% availability post-launch.
  • Cut new pipeline onboarding from 8 weeks to 4 through self-service infrastructure automation and standardized deployment patterns.
#0c3d8e5
Amazon Web Services
Software Development Engineer
Dallas, TX
Jan 2020 — Nov 2020
COMPLETE · 100% uptime
$ deploy --target=amazon-web-services --role="Software Development Engineer"
BUILD TEST DEPLOY
Internal developer tooling
  • Shipped Python SDK abstractions for internal AWS developer tooling used by thousands of engineers.
  • Built client-side permission validation for the object storage SDK that eliminated a class of runtime authorization failures.
#4d2f9b0
Nisum
Software Engineer Intern (Internship)
Fremont, CA
Jun 2018 — Aug 2018
COMPLETE · 100% uptime
$ deploy --target=nisum --role="Software Engineer Intern"
BUILD TEST DEPLOY
E-commerce promotions system — backend servicesJava · Spring Boot
  • Reduced average API response time by 65% by implementing an application-layer caching strategy (EhCache) and resolving key performance bottlenecks in the REST API layer.
  • Optimized the data layer by tuning the JDBC connection pooling strategy, which successfully reduced round-trip database latency by 15ms per request.
  • Ensured system reliability and code quality by writing comprehensive unit tests (JUnit) and conducting performance load testing with JMeter before shipping to production.
#1a9c3f2
Deutsche Bank
Software Developer (Apprenticeship)
Pune/Pimpri-Chinchwad Area
Jan 2015 — Dec 2015
COMPLETE · 100% uptime
$ deploy --target=deutsche-bank --role="Software Developer"
BUILD TEST DEPLOY
QA automation frameworks — banking applicationsJava · Selenium · HP UFT
  • Improved overall testing efficiency by 50% by developing a test automation framework from the ground up using Java and Selenium, significantly reducing manual regression testing hours.
  • Ensured data integrity for critical financial transactions by performing rigorous backend database testing, writing complex SQL queries to validate data transformations and identify anomalies.
  • Engineered automated test suites for both modern web (Selenium) and legacy desktop applications (HP UFT), establishing comprehensive test coverage across the bank's technology stack.
cat SKILLS.md

Skills

Distributed Systems & Backend
GogRPC / ProtobufTemporalMicroservicesAPI DesignConcurrencyDistributed LockingIdempotent APIsDeterministic ReplayMulti-region / Multi-colo
Reliability & Production Engineering
Deployment & Change SafetySLI/SLO ManagementError BudgetsCapacity PlanningIncident ResponseOn-callPostmortemsBlast Radius ContainmentProgressive DeliveryFail-open Design
Infrastructure & Observability
LinuxKubernetesDockerTerraformPrometheusGrafanaOpenTelemetryDistributed TracingKQL / Kusto
AI for Reliability
Agentic TriageEvidence-grounded RAGLLM Tool IntegrationGenAI Guardrails
Languages
GoPythonJavaJavaScriptBash
cat EDUCATION.md

Education

M.S. Computer Science
Santa Clara University
2019
B.S. Computer Science
Pune University
2016