nikhil@platform:~$
Nikhil Vaidyar

Nikhil Vaidyar

Senior Platform Engineer. Go, Kafka and Kubernetes on AWS. Six years keeping peak e-commerce and health-tech traffic fast and correct, plus merged work in Kubernetes and Argo CD. Based in Bengaluru, looking for a platform team in the Netherlands, Ireland or Germany.

$ mail nikhil resume.pdf github/nikzayn linkedin
nikhil@platform: ~
kubectl describe engineer nikhil-vaidyar
Name:
Nikhil Vaidyar
Role:
Senior Platform Engineer
Location:
Bengaluru, India (UTC+5:30)
Status:
Running open to roles in NL, IE, DE; needs visa sponsorship
Labels:
go, kafka, kubernetes, eks, aws, terraform, argo-cd, sre
Certs:
CKAD, Harness CD & GitOps Developer
Containers:
victorias-secretSenior SWE, PlatformRunning 2025-03
xaiSWE, AI Tutor, remoteRunning 2026-05
Fleet:
8 EKS clusters, 120+ nodes, 2,000+ pods, 120+ services on GitOps
Metrics:
peak_rps20,000+
read_p99_ms850 → 220
mttr_minutes35 → 10
oversell_rate1–2% → ~0%
$

journey

Tap a revision to expand it.

rollout history
kubectl rollout history deployment/nikhil-vaidyar
  1. #10May 2026 – nowSWE, AI Tutor · xAI, remotedetailsLIVE
    • Contribute to AI model training by curating high-quality code samples, validating generated output and delivering precise corrections across specialised programming languages, improving response accuracy and code reliability.
    • Evaluate and refine AI-generated code for scalability, efficiency and maintainability: finding logical inconsistencies, optimising implementations and aligning output with industry engineering practice.
    • Work with cross-functional engineering and AI teams to raise quality benchmarks for enterprise-grade AI coding, with more consistent performance and reliable developer-focused output.
  2. #9Mar 2025 – nowSenior Software Engineer, Platform · Victoria's Secret & CodetailsLIVE

    I own the delivery, caching, streaming and observability paths behind peak e-commerce traffic.

    • Designed and shipped the flash-sale order and inventory platform in Go on EKS, with Redis-backed reservations (TTL, idempotency keys, atomic operations). 20K+ RPS and 1,200+ orders a minute at peak; overselling from 1–2% to effectively zero.
    • Ended cache stampedes by replacing request-time fan-out with a Kafka precompute pipeline partitioned by shop_id, with bounded worker pools, atomic pointer reads and batched MGET. Peak read p99 850 → 220 ms.
    • Built an AI-assisted SRE platform: Go services and Kubernetes operators instrumented with Prometheus, Grafana, OpenTelemetry and Loki, with an LLM agent doing RAG over runbooks and past incidents to diagnose and auto-run safe fixes. MTTR about 35 → 10 minutes across a 15K+ RPS fleet.
    • HPA and KEDA autoscaling on CPU and Kafka consumer lag (10 → 40 pods in 5 minutes), with rate limiting, circuit breakers, retries and backpressure.
    • Multi-AZ zero-downtime releases, and load-testing that makes event readiness a measured gate instead of a judgement call.
    • Natural-language layer over Prometheus with MCP: plain-English questions become PromQL and service-health scores.
    • Hardened the platform with RBAC, network policies, secrets management and image scanning in CI; alerting through Slack and PagerDuty.
    • Architected VogueFlow, a real-time multi-agent AI commerce engine in Go (LangChain, Claude API, Pinecone, Redis) with tiered semantic caching and model cascading. 15K RPS at sub-200 ms p95, inference cost down 42%.
    • Cut EKS spend 30% (about $18K a month) with right-sizing, Karpenter spot nodes and off-peak scale-down. I also lead design reviews and knowledge-sharing sessions.
  3. #8Apr 2023 – Mar 2025Software Engineer · Victoria's Secret & Codetailsdone
    • Led the migration of 120+ microservices from Consul to Harness and Argo CD GitOps on Helm and GitHub Actions, with environment promotion, pod disruption budgets and automated rollback. 425 releases across 2,500 PRs; deploys 8 → 25 a week.
    • Prometheus and Grafana dashboards that cut on-call troubleshooting by about half.
    • Reworked database access and removed redundant Kafka retry loops: p95 down 30% under peak load.
    • Rebuilt Add to Bag for Beauty and Sleepwear: +$2.3M annualised sales.
  4. #7Feb 2023 – Oct 2024Upstream contributor, CNCF · Kubernetes, Argo CDdetailsmerged
    • Kubernetes contributor (Feb 2023 – Apr 2024): safe sysctls change that ships in v1.32.
    • Argo CD contributor (May – Oct 2024): two PRs that ship in v2.12.
    • cloud-provider-azure: registry migration and docs fix. See the PRs.
  5. #6Feb – Jun 2023Computer science program · Scalerdetailsdone
    • Deepened CS fundamentals; certified CKAD and Harness CD & GitOps Developer.
  6. #5May 2020 – Mar 2023Software Development Engineer · Biofourmisdetailsdone
    • Event-driven ingestion in Go, Kafka and AWS Lambda over PostgreSQL and MongoDB for biometric telemetry from 100K+ remote patients: 400B+ data points a day, 5x throughput, clinical alert latency down 40%.
    • Configurable care-workflow engine for 50+ clinical protocols, p95 under 250 ms across 100K+ concurrent sessions.
    • Built and operated the microservices behind continuous biometric collection in a regulated environment.
  7. #4Jan – Jun 2021Volunteer · Knaxusdetailsdone
    • Helped run a community where developers collaborate, learn and grow.
  8. #3Dec 2018 – Jan 2020Full-stack developer intern · JoosWorks, EXLdetailsdone
    • JoosWorks (Dec 2018 – Jan 2020): Shopify SKU automation on AWS Lambda and analytics dashboards.
    • EXL (Mar – Sep 2019): features for the MIA business-intelligence tool in Go, MySQL, React and Docker. My first Go in production.
  9. #2Sep 2017 – Oct 2018Web developer · Imxplorer, Corporate Flyersdetailsdone
    • Imxplorer (Sep 2017 – Feb 2018): tourism and events booking site, including its SEO.
    • Corporate Flyers (Mar – Oct 2018): web development alongside the design team.
  10. #1Aug 2015 – May 2019B.Tech, Electronics & Communication · Rajasthan Technical Universitydetailsdone
    • Signals, circuits and systems, then software took over.

impact

Before and after, in production.

kubectl top
kubectl top engineer nikhil-vaidyar --compare=before
read p99Kafka precompute replaced fan-out
before850 ms after220 ms
mttrobservability + LLM agent over runbooks
before35 min after10 min
deploys / week120+ services on GitOps
before8 after25
oversold ordersRedis reservations, idempotency keys
before1–2% after~0%
eks spendright-sizing, Karpenter spot
before100% after70%
llm inference costsemantic caching, model cascading
before100% after58%

open source

Merged and shipping in tagged releases.

gh
gh pr list --author nikzayn --state merged
MERGEDAdd net.ipv4.tcp_rmem and tcp_wmem to the kubelet's safe sysctlskubernetes/kubernetes · via #127489, issue #125234v1.32
MERGEDAdd domain_hint as an optional OIDC config fieldargoproj/argo-cd · #18214v2.12
MERGEDCLI hint for getting extended outputargoproj/argo-cd · #18253v2.12
MERGEDMove image references from k8s.gcr.io to registry.k8s.iokubernetes-sigs/cloud-provider-azure · #3324, plus docs fix #35802023

projects

$ ls ~/projects

shop-flow/IN PROGRESS

go, aws, terraform, postgres, sqs, keda, grafana-lgtm, otel

A flash-sale-ready e-commerce platform built from zero: transactional outbox to SQS, KEDA-scaled workers, SLOs with error budgets, self-rolling-back canaries and a full game-day sale simulation.

golang-banking/

go, kubernetes, helm, grpc, jwt, sqlc

Domain-driven banking service with pessimistic locking on transaction paths, so balances stay consistent under concurrent transfers.

image-service/

python, api-gateway, lambda, s3, dynamodb

Uploads go straight to S3 through signed forms with a pending-to-active lifecycle, so bytes never touch Lambda.

golang-lld/

go

Low-level designs for parking lots, food delivery, movie booking and more, for design-interview prep.

sqlite-db-go/

go

A SQLite clone, to learn how a database stores and finds rows on disk.

stack & writing

skills.yaml
cat skills.yaml
cloud: [aws, eks, lambda, s3, iam, vpc, rds, dynamodb, sqs, gcp, aks]
iac: [terraform]
kubernetes: [ckad, helm, argo-cd, harness, keda, hpa, karpenter, rbac]
reliability: [prometheus, grafana, opentelemetry, loki, slos, pagerduty]
data: [kafka, redis, postgres, mysql, mongodb, elasticsearch, grpc]
languages: [go, python, java, sql, bash]
ai: [langchain, langgraph, rag, mcp, claude-api, pinecone]
certs: [ckad, harness-cd-gitops, lfd102]
spoken: [english, hindi]  # learning german + dutch
medium
ls ~/writing
  • 2024
    Getting started with KyvernoLevel Up Coding
  • 2023
    Data structures in GolangLevel Up Coding
  • 2022
    Principles of OOP in GoBetter Programming
  • 2022
    The fundamentals of pointers in GoBetter Programming
  • 2021
    Golang basics: maps, arrays, slicesnikhilvaidyar.medium.com
open medium ↗

contact

Email is quickest.

mail
./contact --hire nikhil-vaidyar
email:
nikhilvaidyar1997@gmail.com
phone:
+91 70422 24751
linkedin:
in/nikhil-vaidyar
looking_for:
Platform, SRE or infrastructure roles
regions:
Netherlands, Ireland, Germany (open to relocate)
visa:
Needs sponsorship # eligible: EU Blue Card, Dutch HSM, Irish Critical Skills
timezone:
UTC+5:30, happy to overlap with European hours
Send an email Call Download resume