Open to staff+ IC, leadership & consulting
Saurabh Singh — Backend & Platform EngineerEngineering platforms that stay fast, reliable, and operable at scale.I build high-throughput platform systems - modernizing monoliths, orchestrating workflows, and leading teams that ship reliably at scale.
01What I solve
Three problem shapes I am most often brought in for — each linked to the work that evidences it.
Break apart tightly coupled systems without disrupting the clients that depend on them.
24+ modules decomposed, 0 breaking API changes
Build event-driven architectures that hold up under high-throughput, high-value workloads.
16s → <1s execution across 2M+ transactions
Raise deployment reliability, observability, and operational efficiency on production cloud platforms.
HSM-backed custody for 100K customers
02Selected work
Scroll through each one: how it was shaped, what changed, and what it measured.
01 Transaction platform 2025
Decomposed a 24+ module transaction monorepo into independently deployable services while preserving external API contracts and workflow orchestration.
02 Router Protocol 2024-2025
Reduced critical transaction execution from 16 seconds to under 1 second while handling 2M+ transactions and $1B+ in volume.
03 CoinDCX 2022-2024
Designed and delivered a cold-custody wallet platform with HSM-protected keys, multi-signature authorization, and a sub-30ms configurable policy engine. Served 100K customers and 10K daily transactions with full auditability and no per-rule code deploys.
03Operating principles
If support and QA cannot reason about a system, it is not done.
Prefer incremental migration over big-bang rewrites; preserve external contracts while improving internals.
Invest in automation that reduces toil: CI/CD, workflow test frameworks, and actionable observability.
Mentor through ownership - give engineers end-to-end problems, not ticket queues.
04Writing
Notes on platform modernization, coding-agent tooling, and technical leadership.
A function that only reads a status code was secretly opening a new TCP and TLS connection on every call. Here is why HTTP/1 made that unavoidable through Go 1.26, what Go 1.27 changed, and a benchmark that proves it.
Build a Go operator with Operator SDK that rolls a Deployment to each new GitHub release without fighting Helm or KEDA. Tested on kind, failures included.
A hands-on tutorial: send a message between two servers so that only a specific, measured enclave on the receiving side can ever read it, using AES-GCM, RSA keys in KMS, and an attestation document.
Next step
Ten years of backend and platform engineering across Java, Kafka, Redis, Camunda, and AWS — modernizing monoliths, building event-driven systems, reducing operational noise, and leading teams through delivery.
Architecture reviews, embedded delivery leadership, or ongoing advisory — scoped to the problem, grounded in production experience.