Hi I'm Sherod

Platform Engineering Manager @ Bloomberg

New York, NY
Open for new opportunities
Sherod Taylor

Platform Engineering Manager at Bloomberg, leading two teams (10 engineers) building AI tooling and the developer platform 9,000+ engineers ship on every day. 12+ years turning infrastructure into leverage for other teams — most recently making internal platform APIs callable by AI agents end-to-end. Off-hours: home lab of three Claude Code agents that open and review their own PRs.

AI Platforms
Distributed Systems
API Architecture
Full-Stack
Leadership
Project Management
Product
Design

→ Skills & Expertise

Technologies and tools I use to bring ideas to life

  • Frontend Development

    Frontend Development

    12+ years

    • Building modern, responsive web applications

    ReactReact
    AngularAngular
    Next.jsNext.js
    TypeScriptTypeScript
    JavaScriptJavaScript
    Tailwind CSSTailwind CSS
    Vite
    WebpackWebpack
    Module Federation
  • Backend Development

    Backend Development

    12+ years

    • Scalable server-side architecture and APIs

    Node.jsNode.js
    TypeScriptTypeScript
    GoGo
    PythonPython
    PostgreSQLPostgreSQL
    RedisRedis
    GraphQLGraphQL
    Apache KafkaApache Kafka
    Apache AirflowApache Airflow
    Event Sourcing
    API Integration
  • DevOps & Infrastructure

    DevOps & Infrastructure

    10+ years

    • Cloud infrastructure and deployment automation

    DockerDocker
    GitGit
    TerraformTerraform
    PackerPacker
    ProxmoxProxmox
    OpenStackOpenStack
    Networking
    TrueNASTrueNAS
    KubernetesKubernetes
  • Leadership & Strategy

    4+ years

    • Team leadership, product management, and design thinking

    Leadership
    People Management
    Project Management
    Product Strategy
    Design Thinking
    Mentoring

→ Work Experience

13+ years of building scalable software solutions

  • Bloomberg LP logo

    Platform Engineering Manager

    Bloomberg LP
    March 2023 - Present
    New York, NY

    • Platform Engineering Manager leading two teams (10 engineers) building AI tooling and the developer platform that powers 9,000+ Bloomberg engineers. Focused on making internal platform APIs callable by AI agents end-to-end alongside the human users they already serve.

    Key Achievements

    • Lead two engineering teams (10 engineers total) covering AI platform tooling and the internal developer platform
    • Drove the bridge from the team's API surface to agentic developer tools — internal libraries, gateway changes, and identity work so AI agents authenticate as first-class clients alongside humans
    • Founded the team's microfrontend framework (Vite + Module Federation): other product teams self-serve onto the platform with full auth, deployment, and shared UI primitives — collapsing a former central-team bottleneck
    • Coach engineers to work ahead via POCs in the margin: validate a direction, surface real constraints, graduate the work into the roadmap with known tradeoffs
    • Stay hands-on in the codebase — model the working-ahead behavior I expect from the team and run POCs alongside ICs

    Tech Stack

    Go iconGo
    TypeScript iconTypeScript
    React iconReact
    Module Federation iconModule Federation
    Kubernetes iconKubernetes
    Observability iconObservability
    Python iconPython
    AI / LLM tooling iconAI / LLM tooling
  • Bloomberg LP logo

    Senior Software Engineer

    Bloomberg LP
    August 2018 - March 2023
    New York, NY

    • Founding engineer on Bloomberg's internal infrastructure provisioning platform. Architected full-stack software for the platform from day one and grew into a tech-lead role while staying hands-on in the codebase.

    Key Achievements

    • Founding engineer on the internal infrastructure provisioning platform — designed the allocation lifecycle state machine, API-level observability instrumentation, and the workflow analytics stack used to drive SLO definition and capacity planning
    • Built the platform's unified API gateway as the external API surface for managed infrastructure — single consistent contract regardless of which underlying system serves the request
    • Designed and shipped the API-level observability layer (per-endpoint metrics, per-team SLO dashboards, runtime-switchable) before the team had any API observability
    • Grew into tech-lead role while staying hands-on — modeled the working-ahead instinct that became how the team operates

    Tech Stack

    Python iconPython
    Go iconGo
    JavaScript iconJavaScript
    React iconReact
    Node.js iconNode.js
    Module Federation iconModule Federation
    GraphQL iconGraphQL
    Observability iconObservability
    Kubernetes iconKubernetes
  • Paxos (formerly itBit) logo

    Senior Software Engineer

    Paxos (formerly itBit)
    April 2015 - August 2018
    New York, NY

    • Led infrastructure modernization and REST API platform development. Designed SWIFT banking system integration and implemented zero downtime deployments.

    Key Achievements

    • Completed full Terraform migration of manually deployed systems
    • Designed and implemented TLS client certificate authentication
    • Led REST API platform development for institutional trading
    • Designed SWIFT banking system for institutional trading integration
    • Implemented zero downtime deployments using HA Proxy
    • Proposed framework using Node.js, React.js, and Flux architecture
    • Contributed to becoming second-largest Bitcoin exchange by volume

    Tech Stack

    Node.js iconNode.js
    React.js iconReact.js
    Terraform iconTerraform
    Docker iconDocker
    PostgreSQL iconPostgreSQL
  • Madison Reed logo

    Software Engineer

    Madison Reed
    September 2013 - May 2015
    San Francisco, CA

    • Implemented Redis caching system, developed custom CMS, and led migration from Magento to custom Node.js framework.

    Key Achievements

    • Implemented Redis caching system reducing database calls
    • Developed custom CMS tailored to company workflow
    • Led migration from Magento to custom Node.js order processing framework

    Tech Stack

    Node.js iconNode.js
    Redis iconRedis
    MongoDB iconMongoDB
    JavaScript iconJavaScript
    CSS iconCSS

→ Personal Projects

Technical projects showcasing AI infrastructure, development tools, and automation solutions

agent-smith
• Autonomous AI engineering teammates that live inside the home lab cluster and treat Matrix as the input channel. Tag a bot, it opens a PR; tag the other bot, it reviews; the product-manager bot owns PRDs and roadmap. They auto-address review comments, use the cluster's own observability stack via MCP, and never hold the real credentials — those stay behind an egress proxy at the cluster edge. Packaged as a public Helm chart anyone can deploy.
Three-Agent Crew, One Parametric Image:DevBot (code), InfraBot (k3s/Flux), and PMBot (product/PRDs) run as separate StatefulSets from the same ghcr.io/sherodtaylor/agent-smith image. Per-agent persona, MCP config, and subagents live under agents/<name>/; the shared base CLAUDE.md is one ConfigMap
Matrix-Driven, Autonomous PR Workflow:Messages in #dev / #infra are real Claude Code prompts. After opening a PR the author mentions the other bot for review; the reviewer runs the code-review skill and posts inline findings. A Stop-hook re-wakes the author on unaddressed comments so iteration happens without a human in the loop
Egress Credential Firewall (iron-proxy):Productionized so agents never hold real GitHub or Anthropic tokens — iron-proxy MITMs egress with a private CA and swaps placeholder tokens for real credentials per-host against a domain allowlist. A compromised pod leaks nothing useful
MCP for Everything Observable:VictoriaMetrics + VictoriaLogs MCP servers, a stdio NATS MCP server for the durable event log, and the Matrix channel plugin for inputs. 'Check the api-latency dashboard' becomes a single prompt
Two-Pane Runtime:Each pod runs claude in tmux — pane 0 owns the Matrix identity, pane 1 runs a second claude --remote-control with its own $HOME so humans can attach via kubectl exec or the Claude desktop/web app
Public OSS Distribution:image (ghcr.io/sherodtaylor/agent-smith) and Helm chart (oci://ghcr.io/sherodtaylor/charts/agent-smith) ship every release; framework docs at sherodtaylor.github.io/agent-smith
Claude Code
MCP
AI Agents
Matrix
NATS
k3sk3s
Flux GitOps
Helm
iron-proxy
Bun
GoGo
DockerDocker
tmux
GitHub Actions
sherodtaylor/agent-smith
SpareCoins icon
SpareCoins
• A crypto investment mobile application that rounds up spare change from everyday transactions and automatically invests it in cryptocurrency. Think Acorns for crypto - democratizing crypto investing through micro-investing with seamless bank integration.
Automated Crypto Investing:Connects to bank accounts to round up transaction spare change and automatically purchases cryptocurrency through Coinbase exchange integration
GraphQL-First Architecture:Built robust backend API using Golang, GraphQL, and Echo framework with OAuth authentication. Generated type-safe client code using GraphQL codegen for seamless frontend integration
Rich Mobile UX:Developed React Native frontend with Apollo GraphQL for data fetching. Created engaging user experience with custom Lottie animations for transaction flows, account linking, and investment progress visualization
Financial Integration:Implemented secure bank account connectivity using Plaid API for real-time transaction monitoring with automated round-up calculations and crypto purchase execution
React Native
GolangGolang
GraphQLGraphQL
Apollo
Lottie
OAuth
Echo
Plaid
Coinbase API
FinTech
sherodtaylor/sparecoins
Private
Home Lab Cluster
• Self-hosted k3s cluster running every app my family actually depends on. Flux reconciles every manifest from sherodtaylor/homelab, Traefik fronts the lot with auto-renewed Let's Encrypt certs, ExternalSecrets pulls credentials from Infisical, and VictoriaMetrics + VictoriaLogs catch anything that drifts — so the house keeps running while I keep tinkering.
Media + Entertainment:Jellyfin and Plex stream the family library, Jellyseerr handles requests, the *arr stack (Sonarr, Radarr, Lidarr, Prowlarr) automates discovery, qBittorrent and Sabnzbd handle downloads, Audiobookshelf catalogues the audiobook collection
Photos + Files + Storage:Immich runs face recognition over years of family photos, Nextcloud hosts personal files, TrueNAS (8 TB via NFS) backs every stateful service in the cluster
Home Automation:Home Assistant runs the house — lights, thermostats, schedules, sensors across the home — all on the same cluster, not a separate appliance
Observability Built-in:VictoriaMetrics + VictoriaLogs ingest every pod's metrics and logs; Grafana dashboards and Alertmanager catch regressions before they become outages
GitOps All the Way Down:Flux reconciles every Kustomization and HelmRelease from sherodtaylor/homelab; cert-manager + kubernetes-replicator handle TLS across namespaces; the cluster rebuilds from a single git push
ProxmoxProxmox
k3sk3s
Flux GitOps
Helm
Traefik
cert-manager
ExternalSecrets
Infisical
TrueNASTrueNAS
Home AssistantHome Assistant
JellyfinJellyfin
Plex
Immich
Nextcloud
Audiobookshelf
Sonarr
Radarr
Prowlarr
qBittorrent
VictoriaMetrics
VictoriaLogs
Grafana
Alertmanager
Let's Encrypt
sherodtaylor/homelab
Dotfiles
• Maintained comprehensive dotfiles and development environment configurations for consistent, reproducible development setups across multiple systems.
Cross-Platform Configuration:Unified development environment setup for macOS and Linux systems
Editor Configuration:Advanced Neovim configurations with language servers and productivity enhancements
Neovim
Zsh
GitGit
Tmux
Shell
macOS
Linux
DockerDocker
sherodtaylor/dotfiles

→ Community Involvement

Giving back through mentorship, volunteer work, and community leadership

  • Bloomberg Diversity Interview Prep Bootcamp

    Mentorship

    Volunteer Instructor & Mentor

    6-month program

    • A comprehensive bootcamp program designed to train underrepresented candidates for technical interviews through mentoring and hands-on training from Bloomberg engineers. Provides 1:1 guidance, curriculum development, and mock interview preparation.

    Key Contributions

    • Designed and delivered Systems Design curriculum
    • Took on 1:1 mentee and prepared them for coding interviews
    • Provided weekly mock interviews and detailed feedback sessions
    • Followed training schedule for 6 months: algos, system design, coding interviews
  • Bloomberg Community Service Leadership

    Volunteer

    Community Service Leader

    Ongoing

    • Leading community service initiatives within Bloomberg engineering teams, organizing volunteer activities and coordinating outreach programs to support local NYC communities.

    Key Contributions

    • Logged 15+ volunteer hours in past year
    • Mobilized multiple engineering teams for community service
    • Coordinated homeless aid initiatives and care package distribution
    • Organized NYC park cleanup efforts

Latest Articles

Thoughts on software development, leadership, and technology