[{"content":"Hey! I\u0026rsquo;m Shubham — a Software Engineer who builds the backend for your backends. Control planes, APIs, workflow engines: the stuff that sits behind the stuff users actually see. It\u0026rsquo;s my job and my passion, and the passion came first.\nI like systems where failure is a first-class feature: idempotency, retries, backpressure, graceful degradation. I like it better when the mistake never gets a chance to fail — the rule that lives in the schema instead of in someone\u0026rsquo;s memory, the test that locks out a class of bug instead of the one instance you just found. I\u0026rsquo;m happiest when a graph goes down because we fixed it, not because we hid it.\nI\u0026rsquo;ve had the chance to speak at DevOpsDays Bengaluru 2023 (because apparently I have opinions about API gateways now), and once won a hackathon by building a blockchain thing before it was cool. Also did a semester at UC Berkeley\u0026rsquo;s Sutardja Center, soaking up the Silicon Valley chaos.\nCurrently obsessed with: policy as schema data instead of code you have to remember to call, invariants that fail at import time rather than in production, and reviewing generated code at its failure paths — that\u0026rsquo;s where it\u0026rsquo;s thin.\nWhat I\u0026rsquo;ve Built @ Lyric.tech Senior Software Engineer, Core (Platform) — 2025–Present\nCore backend services in Python and TypeScript: entity reconciliation, workflow DSLs, Temporal workflows, and the platform and data APIs behind them. Design focus: correctness under change — evolvable schemas, retries that are safe to repeat, and operational clarity (alerts that mean something). I own core platform surfaces across entity reconciliation, workflow definitions, deterministic execution, and the APIs that turn inconsistent customer data into stable internal contracts. Getting a front-row seat to supply chain reality: messy inputs, real-world constraints, and a lot of \u0026ldquo;it depends\u0026rdquo;. @ Tune AI (formerly NimbleBox.ai) Software Engineer — 2020–2025\nThe arc says more than the title does: ML scripts → services I owned end to end → the Go control plane, its Kubernetes operator and the IaC underneath → shared platform libraries → in-process frameworks that shipped with the argument for them written down.\nDesigned the gRPC backbone for model serving, schema-first, so clients and servers could evolve without a flag day. It carried 100M tokens/day and latency dropped 40%. Built document processing on batch inference with replayable runs — a failed batch could be re-run without double work. 100K documents daily, 54% cheaper, precision-recall held at 95%. Wrote a reusable cron scheduler that runs in-process, and shipped it with a \u0026ldquo;Why?\u0026rdquo; README making the case against Kubernetes CronJobs for our workloads. Killed 70% of the boilerplate scheduling code. You\u0026rsquo;re welcome, future me. Led the OpenTelemetry adoption because debugging distributed systems without traces is pain. Shipped a GenAI architecture for a UAE event management company — automated 40% of their workflow and cut proposal creation time by 20%. Where it all started: ML pipeline orchestration, containerized model serving with FastAPI, and liberating models from Jupyter notebooks (not a fan). Side Projects A mix of things I\u0026rsquo;ve built, fixed, or accidentally broke:\ncloud-credit-system — architecture experiment. Credit accounting on TigerBeetle and Temporal. Idempotent by construction: replay a request and the ledger refuses to charge twice. proto-opt-parse — Go package. Runtime parsing of Protobuf descriptor options. Built to make gRPC tooling less guessy. aws-sso-check — maintained CLI. AWS SSO expiry notifier, rewritten from shell to Swift to Go. golang/go#71738 — upstream issue. Pseudo-version stamping in cmd/go doesn\u0026rsquo;t work for a module in a subdirectory. Found it the way you find these things: by not believing a version number. Kubernetes Dashboard #7093 — merged upstream. Found the pod page crashing on generic ephemeral volumes, filed it, then fixed it myself. It was bugging me. Colbert AI — archived team project. Built at Next Tech Lab with two others: GPT-2 (345M) fine-tuned on 500+ transcripts to mimic Stephen Colbert\u0026rsquo;s monologue voice. Still posting. Was doing LLM hijinks before it was fashionable. 😎 Notes, Talks, and Other Escapes DevOpsDays Bengaluru 2023: \u0026ldquo;Using APISIX as a Reverse Proxy\u0026rdquo; — reverse proxying high-volume inference traffic. Co-presented with Vinuja Khatode. Visiting Scholar, UC Berkeley SCET (2019): Crashed the Data-X program on a $20K scholarship and built Docker Hub analytics pipelines. Speaker bio — for anyone who needs one:\nI build the platform layer other engineers build on: control planes, workflow DSLs, entity reconciliation, and schemas that refuse to let you ship a mistake. I talk about API gateways at conferences and still get a kick out of hacking on protocol compilers. Community builder, systems tinkerer, and far too interested in what protobuf descriptors will let you get away with.\nEarlier Next Tech Lab AP, Board Member — helped scale it from fewer than 20 people to 150+ members. ACM Student Chapter, Founding Member — started it from scratch, grew it to 200+ members. Teaching Assistant, SRM University (2019–2020) — ran CS tutorials, built Python grading tools, and helped a lot of people through their first traceback. Winner, IIIT-Delhi HACKIIITD 2017 — 1st of 800+ teams, for a blockchain-based consensus mechanism against misinformation. Yes, blockchain. It was 2017, don\u0026rsquo;t judge. Google Cloud Associate Cloud Engineer · President\u0026rsquo;s Placement Award, SRM University · Grand Finalist, TCS IT Wiz 2013. Tech I Actually Use Daily Drivers Go (control planes, tooling), Python (services, ML infrastructure), TypeScript (platform services) Backend gRPC + protobuf (including custom options), ConnectRPC + buf, FastAPI, Temporal, Apache APISIX Infrastructure Kubernetes (operators, CRDs, Kubebuilder), Docker, AWS, GCP, Terraform, Pulumi Data Stuff Kafka, RabbitMQ, PostgreSQL, Redis, TigerBeetle Observability OpenTelemetry, distributed tracing, structured logging Let\u0026rsquo;s Talk I\u0026rsquo;m happy to talk about platform architecture, workflow systems, Go infrastructure, and conference talks. I\u0026rsquo;m currently employed and not looking for a move, but I\u0026rsquo;m open to speaking invitations and open-source collaboration. I usually reply within a week.\nReading I keep a running shelf log on the Reading page.\nEmail: contact@cshubhamrao.dev GitHub: cshubhamrao LinkedIn: cshubhamrao X: @cshubhamrao ","permalink":"https://cshubhamrao.dev/about/","summary":"\u003cp\u003eHey! I\u0026rsquo;m Shubham — a Software Engineer who builds the backend for your backends. Control planes, APIs, workflow engines: the stuff that sits behind the stuff users actually see. It\u0026rsquo;s my job and my passion, and the passion came first.\u003c/p\u003e\n\u003cp\u003eI like systems where \u003cstrong\u003efailure is a first-class feature\u003c/strong\u003e: idempotency, retries, backpressure, graceful degradation.\nI like it better when the mistake never gets a chance to fail — the rule that lives in the schema instead of in someone\u0026rsquo;s memory, the test that locks out a class of bug instead of the one instance you just found.\nI\u0026rsquo;m happiest when a graph goes down because we fixed it, not because we hid it.\u003c/p\u003e","title":"About Me"},{"content":"Quick shelf so future-me remembers what’s here. Trying to read more fiction without turning this into a review blog.\nStatus key ✅ Done 📖 In progress 💤 On deck 📚 Reference 🔁 Re-reading / Reference Fiction \u0026amp; narrative Dune — Frank Herbert (📖) — slow, savoring the desert politics. Yellowface — R.F. Kuang (✅) — sharp and fast. Tomorrow, and Tomorrow, and Tomorrow — Gabrielle Zevin (✅) — friendship + games. Light reads \u0026amp; meta Anxious People — Fredrik Backman (✅) — small-town chaos. Strange Planet — Nathan W. Pyle (✅) — palate cleanser panels. Chip War — Chris Miller (💤) — parked for a focused weekend. Systems \u0026amp; engineering Designing Data-Intensive Applications — Martin Kleppmann (🔁) — storage sanity checks. Concurrency in Go — Katherine Cox-Buday (🔁) — goroutine compass. The Software Engineering Guidebook — Gergely Orosz (✅) — read on release. Language \u0026amp; reference The Go Programming Language — Donovan \u0026amp; Kernighan (✅) — read deeply. Zero to Production in Rust — Luca Palmieri (📖) — paused midway. Writing an Interpreter / Compiler in Go — Thorsten Ball (📚) — reach-for guides. ","permalink":"https://cshubhamrao.dev/books/","summary":"\u003cp\u003eQuick shelf so future-me remembers what’s here. Trying to read more fiction without turning this into a review blog.\u003c/p\u003e\n\u003ch2 id=\"status-key\"\u003eStatus key\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e✅ Done\u003c/li\u003e\n\u003cli\u003e📖 In progress\u003c/li\u003e\n\u003cli\u003e💤 On deck\u003c/li\u003e\n\u003cli\u003e📚 Reference\u003c/li\u003e\n\u003cli\u003e🔁 Re-reading / Reference\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2 id=\"fiction--narrative\"\u003eFiction \u0026amp; narrative\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eDune — Frank Herbert (📖) — slow, savoring the desert politics.\u003c/li\u003e\n\u003cli\u003eYellowface — R.F. Kuang (✅) — sharp and fast.\u003c/li\u003e\n\u003cli\u003eTomorrow, and Tomorrow, and Tomorrow — Gabrielle Zevin (✅) — friendship + games.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2 id=\"light-reads--meta\"\u003eLight reads \u0026amp; meta\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eAnxious People — Fredrik Backman (✅) — small-town chaos.\u003c/li\u003e\n\u003cli\u003eStrange Planet — Nathan W. Pyle (✅) — palate cleanser panels.\u003c/li\u003e\n\u003cli\u003eChip War — Chris Miller (💤) — parked for a focused weekend.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2 id=\"systems--engineering\"\u003eSystems \u0026amp; engineering\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eDesigning Data-Intensive Applications — Martin Kleppmann (🔁) — storage sanity checks.\u003c/li\u003e\n\u003cli\u003eConcurrency in Go — Katherine Cox-Buday (🔁) — goroutine compass.\u003c/li\u003e\n\u003cli\u003eThe Software Engineering Guidebook — Gergely Orosz (✅) — read on release.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2 id=\"language--reference\"\u003eLanguage \u0026amp; reference\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eThe Go Programming Language — Donovan \u0026amp; Kernighan (✅) — read deeply.\u003c/li\u003e\n\u003cli\u003eZero to Production in Rust — Luca Palmieri (📖) — paused midway.\u003c/li\u003e\n\u003cli\u003eWriting an Interpreter / Compiler in Go — Thorsten Ball (📚) — reach-for guides.\u003c/li\u003e\n\u003c/ul\u003e","title":"Reading"}]