01
Meeteam · Senior Full Stack Developer
Feb 2026 — Present·8 months
Team: 15+ · devs, UX, QA, PO, DevOps
Chapter 1 · Webinar Platform · Integrations
Connecting webinars to the systems that actually close deals.
The situation
Enterprise clients run their webinars on ON24 and CVENT, but the registration, attendance and engagement data they produce is worth little until it lands in Salesforce, Marketo, HubSpot or Dynamics 365, where sales and marketing act on it. I joined the integrations team to make that flow reliable across North America and Europe, and to stop account managers from uploading CSV lists by hand.
What I did
15+integrations live
ON24/CVENT webinar data flowing into Salesforce, Marketo, HubSpot and Eloqua via Tray.io workflows and TypeScript/Node.js/.NET services. 100K+ leads/month at 98% delivery.
8h/weeksaved per client
LinkedIn → ON24 registration pipeline on the LinkedIn Marketing API: webhook challenge-response, leadNotifications lifecycle, dynamic form-field mapping. No more manual CSV uploads.
90%self-service setups
Tray.io Config Wizard selector scripts for ad accounts, lead gen forms and segments. Account managers deploy integrations without engineering.
Also: Pardot end-to-end sync · 10K+ records/month
TypeScriptNode.js.NETTray.ioSalesforceDynamics 365LinkedIn APIOAuth 2.0
What I took with meIntegration work is trust work. Idempotent upserts, retries on 429 and a config wizard someone else can run are what make a pipeline disappear into the background.
02
SUMZ · Lead Senior Full Stack Developer
Oct 2024 — Feb 2026·16 months
Team: 5 devs · 2 UX · 1 PO
Chapter 2 · Financial Services Platform
From a 45K-line monolith to seven services that ship every day.
The situation
A financial platform for BNPL clients was processing payments, loans and accounts on an Elixir/Phoenix monolith that was hard to change and harder to observe: ten production incidents a month, four hours to resolve each one. I led the team that had to keep it running while rebuilding it on GCP.
What I did
99.97%uptime
Architected cloud-native microservices platform on GCP serving 1,000+ customers, migrating Elixir/Phoenix to Node backend with PostgreSQL, deployed via Cloud Run with auto-scaling.
15+deploys/week
CI/CD pipeline on GitHub Actions + Cloud Build, Docker containerization, Terraform IaC — zero-downtime releases.
90%fewer incidents
Dropped from 10/month to <1/month via observability stack (Cloud Monitoring, Prometheus, structured logging). MTTR 4h → 45min.
Also: 45K lines decomposed · 40% faster reviews
Node.jsTypeScriptPostgreSQLGCPCloud RunTerraformVitestDocker
What I took with meDecomposition only pays off once the boring parts are done: CI/CD, infrastructure as code and an observability stack that tells you what broke before a customer does.
03
Tranqui · Lead Senior Full Stack Developer
Aug 2023 — Sep 2024·13 months
Team: IC in 12-person team
Chapter 3 · Financial Agent Platform
First put out the fire. Then pour the foundations.
The situation
Agents managing customer accounts and transactions were working on a Node.js/Express platform that crashed fifteen times a week, with 45% test coverage and a 15-million-row PostgreSQL database nobody wanted to touch. I came in as an individual contributor in a twelve-person team with a clear mandate: make it stable, then make it fast.
What I did
95%crashes eliminated
From 15 critical failures/week to <1/week. Heap profiling, circuit breakers, N+1 query fixes via TypeORM, rate limiting. Saved $45K/month.
45% → 95%test coverage
Jest (2,500+ cases), Supertest API tests, CI blocks PRs below 80%. Reduced production bugs 68% (25/mo → 8/mo).
<300msupdate latency
Real-time dashboard for 200+ concurrent agents with Socket.io + React/TypeScript + MUI, SSE fallback for legacy browsers.
Also: 85% faster queries · SOC 2 passed audit
Node.jsExpressTypeORMPostgreSQLReactSocket.ioJestMaterial-UI
What I took with meStability is measurable. Heap profiles, circuit breakers and a CI gate at 80% coverage turned $45K a month of downtime into a SOC 2 audit passed.
04
Genius Sports · Senior Full Stack Developer
Jun 2021 — Jun 2023·24 months
Team: Cross-functional · 8 engineers
Chapter 4 · Real-Time Sports Betting
2.5 million people on an NFL Sunday, and 650 milliseconds to show them the odds.
The situation
Live betting widgets for the NFL, the NBA, Caesars and FanDuel had to survive Super Bowl traffic spikes and refresh odds in under a second, on the web and in React Native apps, without scaling the infrastructure five times for March Madness. Everything was a latency problem.
What I did
2.5M+concurrent users
NFL Sunday games. Event-driven microservices, React/TS frontend, Node APIs, GraphQL, WebSocket streaming, AWS Lambda odds calc. P99 380ms.
47%bundle size cut
920KB → 487KB gzipped. Code-splitting with React.lazy, dynamic imports, tree-shaking, progressive JPEGs, Akamai CDN. TTI 4.2s → 1.8s.
82%DB load reduction
12K → 2.2K queries/min. Redis Cluster with consistent hashing, cache-aside, warming, pub/sub invalidation. Handled 5x March Madness growth.
Also: <650ms end-to-end latency · 3.8 → 4.6★ app rating · 95% test coverage
ReactTypeScriptGraphQLNode.jsAWS LambdaRedisWebSocketsReact Native
What I took with meAt that scale every layer counts: GraphQL with DataLoader, Redis Cluster with cache warming, WebSocket pooling, and a test suite large enough to ship thirty times a month without fear.
05
Intergrupo · Junior Developer → Solution Architect
Mar 2013 — Jun 2021·8 years 3 months
Chapter 5 · Consulting · EPM Utilities
Eight years, one company, from junior developer to solution architect.
The situation
I started at Intergrupo writing .NET for EPM, one of Colombia's largest utilities, and stayed through fifteen-plus enterprise projects. The platform I ended up architecting on .NET Core, Azure Functions and SQL Server served 2.5 million utility customers, and the Azure migration behind it saved more than $2M a year.
What I did
$2M+annual savings
Azure cloud migration and process automation across 15+ enterprise projects.
2.5M+utility customers
Architected microservices platform with .NET Core APIs, Azure Functions, SQL Server. Downtime 12h/mo → 2h/mo.
15+devs mentored
Established code review practices, technical design doc standards, weekly architecture workshops.
.NET CoreAzureSQL ServerAzure Functions
What I took with meThis is where I learned that architecture is only half the job. The other half is the fifteen junior developers you leave better than you found them.
