Facts Mode

Victor Sarmiento — infrastructure-minded engineer who ships

Backend systems · search/workflows · ranking logic · reliability habits

CORE SKILLS

Languages: TypeScript/JavaScript · Python · SQL

Web: Next.js · React · Tailwind · Node.js

Backend: REST APIs · Auth/JWT · Supabase RPC · Data modeling

DB: PostgreSQL · Views · Query abstractions

Systems: Ranking logic · Search intent routing · Debugging/RCA · CI/CD (GitHub → Vercel)

Ranking logic + search workflow design for an affiliate discovery platform.

Tech: Next.js · TypeScript · Tailwind · Supabase (RPC + Views) · Vercel

Implemented weighted ranking logic balancing monetization constraints with recency scoring
Encapsulated selection and ranking logic via Supabase RPC; modeled DB views for clean abstractions
Improved search workflow by classifying intent (direct URL vs similarity lookup)
In progress: background search execution to reduce perceived latency

Workflow-driven ticketing system built from real operational pain points.

Tech: Node/Express · React/Vue · PostgreSQL · Auth · Deployment

Designed a ticket workflow focused on reliability, ownership, and operational clarity
Implemented authentication and role-based behaviors for different users
Modeled relational data and optimized queries for common operational views
Shipped an end-to-end product experience (create → triage → resolve → report)

Dual-mode portfolio: Facts for recruiters, Story for narrative context.

Tech: Next.js · TypeScript · Tailwind · Vercel

Designed a split landing experience optimized for both skimmers and deep readers
Built reusable components and consistent visual identity across modes
Focused on clarity, accessibility, and mobile-first behavior

Time-series numerical integration + noise handling for 1000 Hz biomechanical datasets.

Tech: Time-series · Numerical integration · Noise handling · Visualization workflows

Implemented numerical integration to derive velocity/position from acceleration signals
Handled noise/smoothing to stabilize computed series and improve interpretability
Assisted with complexity evaluation and automated visualization outputs

Next.js · Tailwind · Supabase (RPC + Views) · Vercel

Implemented weighted ranking logic balancing monetization constraints with recency scoring
Encapsulated selection/ranking logic via Supabase RPC + DB views
Improved search flow by classifying intent (direct URL vs similarity lookup)
In progress: background search execution to reduce perceived latency
Shipped via PR-based Vercel preview deployments; mobile-responsive UI components

Software Engineer (2019–2021) — Rails · React · PostgreSQL · AWS

Built automation pipeline (PDF → structured mappings → S3) reducing processing from days to hours
Standardized form taxonomy to reduce mapping time and improve maintainability
Shipped behind feature flags; debugged launch blockers and cross-browser issues
Authored 20+ RCAs; monitored production via Airbrake

Workforce Management — 24/7 operational systems

Operate within real-time constraint systems (coverage, compliance, fairness)
Escalations + reliability mindset: what breaks, why it breaks, how to prevent repeats
Translate business rules into consistent, auditable workflow behavior

Numerical modeling · noise handling · performance trade-offs

Implemented numerical integration to derive velocity/position from acceleration signals
Handled noise/smoothing to stabilize computed series
Worked with high-rate (1000 Hz) time-series data and multi-trial aggregation
Assisted with visualization workflow + complexity evaluation