Skip to main content
ssilvestri

Steven Silvestri · Principal Software Engineer

I design the systems other teams build on.

gRPC microservices, event-driven pipelines, and the shared NestJS libraries that product teams build on.

latest post · A Clean AI Review Is Weak Evidence →

About

I build the libraries and services that let other teams ship without thinking about the plumbing.

Most of my work is backend platform engineering: gRPC microservices in Node.js, event-driven pipelines, and the developer experience around them. That means shared abstractions, conventions, and review guidance that keep a fleet of services consistent as more teams build on it.

Lately that includes AI agents in the review and authoring loop, and being honest about what they catch and what they miss.

focus
distributed systems, developer experience
stack
TypeScript, Node.js, NestJS, gRPC
writes about
ai, bdd, redis, kafka

Experience

Where I've worked

Principal Software Engineer

TruMid•Remote
Jan 2018 - Present

Principal Software Engineer developing highly scalable gRPC microservices and managing developer experience for Node.js and NestJS ecosystems

key achievements

  • Develop highly scalable gRPC microservices in Node.js
  • Manage Developer Experience around Node and NestJS by creating libraries and abstractions
  • Provide best practices and guidance for TypeScript and RxJS implementation
  • Tailor AI agents and tools to dramatically increase team productivity and code quality
  • Built Risk Asset Management system tracking real-time PNL and trading metrics

technologies

TypeScriptNode.jsNestJSgRPCRxJSMicroservicesProtocol BuffersMongoDBAI Tools

Software Engineer

TuringsCraft Inc.•Brooklyn, New York
Feb 2016 - Jan 2018

Lead FrontEnd Developer for application redesign of core product using React and Redux

key achievements

  • Lead FrontEnd Developer for core product redesign
  • Drove technology decisions and code quality standards
  • Implemented React and Redux architecture

technologies

ReactReduxES6JavaScriptFlux

Professional Services Engineer

Digital Reasoning•New York, New York
Aug 2015 - Feb 2016

Professional Services Engineer working with machine learning and data analytics, processing unstructured data at scale

key achievements

  • Worked with clients to process unstructured data and meet their goals
  • Wrote ETL code in Groovy and Python for large-scale data processing
  • Optimized code performance for efficient data processing
  • Educated clients on software usage and best practices

technologies

GroovyPythonMachine LearningData AnalyticsETL

Systems

Architecture & systems thinking

As AI writes more of the code itself, what matters is understanding the systems it runs in.

ClientGatewayService AService BService C⚠ downData

01 / 05 · how systems are shaped

Distributed Systems & Service Architecture

Designing how independent services find each other, communicate, and keep working when one of them goes down — decomposition, gateways, and failover, not just the RPC framework underneath.

Service decompositionAPI gatewaysgRPC / service-to-service commsFault tolerance & failoverLoad balancing
PublisherEvent BusSubSubSub

02 / 05 · how services talk

Event-Driven & Asynchronous Systems

Decoupling producers from consumers so each can fail, scale, and deploy independently — message buses and pub-sub over direct, synchronous calls.

Message / event busesPub-sub patternsReactive streams (RxJS)Eventual consistencyDecoupled communication
CommitBuildTest✕ failedDeploy

03 / 05 · how they ship

Deployment & Delivery Architecture

Getting code from commit to running safely, repeatedly, without hand-holding — pipelines and gates that actually gate, not just green checkmarks.

CI/CD pipelinesContainerization & orchestrationInfrastructure as codeBuild systemsAutomated release gates
RequestCacheStoreResponse

04 / 05 · how they keep data

Data Architecture & Storage

Choosing what's fast, what's durable, and how they stay consistent with each other — schema, caching, and read/write tradeoffs, not just which database.

Schema designCaching strategyData pipelines / ETLRead/write path tradeoffsPolyglot persistence
YouAI AgentAI AgentAI AgentReviewProduction System

05 / 05 · where AI fits in

AI-Augmented Systems Engineering

Directing multiple AI agents inside real architectural constraints, and reviewing what they produce for runtime correctness — not syntax. As AI writes more of the code itself, this is the skill that decides whether what it writes actually holds up in production.

Agent orchestrationArchitectural review of generated codePrompt & context engineeringRuntime reasoning over syntax

Projects

Things I've built

all repos on GitHub →

spotlight · open source

Claude Code Doggos

A pixel-art golden retriever that keeps you company in Claude Code. It sniffs the files Claude reads, digs when it edits, tilts its head when Claude asks permission, and falls asleep when you wander off.

typescript · python · claude code plugins

view on GitHub →

Contact

Find me on GitHub or LinkedIn.

Or subscribe to the RSS feed to get new posts.