Lead Backend Engineer with 5+ years building production systems. I design distributed backends and own features end to end, from the gRPC microservices and API gateway down to the Kubernetes and CI/CD that run them. What I care about most is how a system behaves when something fails: idempotency, compensation, integrations that recover. I also build a lot on the side to keep learning across the stack.
By the numbers · a platform I'm a primary contributor to
0+
Years shipping production
0K+
Users served in production
0K+
Transactions handled
~0
Microservices contributed
0+
Side projects built
0
Languages in rotation
STEP 02Experience
Owning systems end to end, from design to production.
From requirements and system design through implementation, testing, and production deployment.
Jul 2023 — PresentPresent
Zazpay
Lead Backend Engineer
Full-time · Mexico City (Hybrid)
Lead backend engineer at a consumer-lending fintech (personal loans, BNPL, and store credit). I lead a backend team and own features from requirements and system design through implementation, testing, and production deployment.
Own personal loans end to end (the company's highest-revenue product), from system design through implementation, testing, and production deployment.
Own the BNPL product end to end: credit lines, confirmation and one-shot payment flows, and dynamic plans.
Built and owned Sunflow, the multi-channel messaging product driving collections (cobranza) over WhatsApp, SMS, email, and voice.
Lead the AI-model integration with the data team, wiring embedding- and LLM-powered features into the platform.
Build fault-tolerant money movement (idempotency, timeouts, circuit breakers, compensation) so payments come out right even when a step fails.
Built CI/CD across 20+ services in both GitLab CI and GitHub Actions, and contribute to GitOps delivery with ArgoCD on GKE.
Instrumented the platform with Datadog (APM, logs, traces) and contribute to the cloud infrastructure managed with Terraform.
Contract software engineering and DevOps: I designed, developed, and deployed products with AI integrations for client companies like Pied and Happymile.
Replaced slow manual Excel reconciliation in a cross-border (US/MX) logistics operation with custom software: a Python/FastAPI engine (pandas, MongoDB) that ingests billing and trip spreadsheets, auto-classifies freight routes, and computes extra-kilometer and toll costs, behind a React/TypeScript operator app on Cloud Run + Firebase.
Contributed to a multi-region (US/MX) Transportation Management System, on FastAPI + DynamoDB and React, deployed on AWS ECS Fargate + Amplify with Terraform and GitHub Actions.
Designed the requirements for an AI-assisted operations CRM (FastAPI + MongoDB, React 19) with OpenAI-powered task auto-classification, auto-assignment, and insights, and led a team of developers delivering it on GCP Cloud Run + Firebase.
An open-source dashboard for BullMQ queues that drops into a NestJS app in one line. It uses NestJS dependency injection to auto-discover your queues and show the processors, workers, and event handlers behind them. Published on npm with zero runtime dependencies.
One-line setup that auto-discovers queues and maps them to their processor classes
An open-source contribution to Bruno, the Git-friendly API client. PR #8374 makes multi-file environment import resilient and idempotent: each file is parsed independently with per-file format detection, failures are isolated instead of aborting the whole batch, and re-importing the same files no longer stacks duplicates.
Per-file parse with isolated error handling — one malformed file no longer aborts the whole import
Idempotent re-import via duplicate detection on sanitized environment names
Summary toast (imported / skipped / failed) backed by 9 new Jest test cases
A web app to format, preview, and share developer documents. It renders Markdown, JSON, Mermaid diagrams, HTML, and plain text in the browser — with KaTeX math and syntax-highlighted code — and turns any document into a shareable link. It also ships an MCP server so AI agents can format and share docs directly.
Formats and previews Markdown, JSON, Mermaid, HTML, and plain text in the browser
Renders Mermaid diagrams, KaTeX math, and syntax-highlighted code
Turns any document into a shareable link; ships an MCP server for AI agents
ReactViteMarkdownMermaidKaTeX
AI / ML★ Featured
Sunflow Agent
Conversational-AI brain for support & collections
The AI agent behind a multi-channel customer-support and collections platform built around the BNPL product. A LangGraph state machine orchestrates embedding-based intent classification, RAG over a Qdrant vector store, and tool-calling flows that hit live service APIs.
LangGraph state machine routing intent → RAG retrieval → tool calls against live APIs
Pluggable LLM provider layer (self-hosted / Ollama with an OpenAI fallback)
Embedding-based intent classification with OpenAI text-embedding-3
PythonFastAPILangGraphQdrantOpenAIRAG
AI / ML
AI Video Pipeline
End-to-end AI YouTube production engine
A 7-stage async orchestrator that turns a topic into a finished video. Pluggable multi-provider stack (LLMs, image gen, AI video, TTS, STT) swappable via one-line YAML, with Remotion + FFmpeg rendering and resumable per-task checkpoints.
Dual workflows: long-form documentaries and zero-LLM 9:16 social shorts
Auto-dubbing into 6+ languages with word-level subtitle timing (Whisper)
PythonRemotionFFmpegClaudeGPT-4oElevenLabs
Market Data
Mayweather
Real-time prediction-markets data platform
Ingests Kalshi and Polymarket order books in real time. Async REST bulk-loaders plus persistent WebSocket streams feed an upsert-based sync into PostgreSQL time-series storage.
RSA-PSS WebSocket auth (Kalshi) + upsert-based sync into PostgreSQL
Orderbook & price-tick time series with a 15-minute catalog refresh
Deployed as a Linux systemd service via GitLab CI/CD with auto-restart
PythonFastAPISQLAlchemyPostgreSQLWebSockets
Sports Tech
PadelRank
Real-time padel ranking & matchmaking app
A React Native + Expo mobile app for points-based ranking, matchmaking, and tournaments, with real-time updates and end-to-end-encrypted messaging over an offline queue.
Dynamic points-based ranking, matchmaking, and event registration
End-to-end encrypted messaging (Signal Protocol) with an offline queue
Cross-platform iOS / Android / web builds via Expo + EAS Build
React NativeExpoSocket.ioSignal ProtocolEAS
Systems
Rust Git TUI
Terminal Git client in Rust (in progress)
A terminal-based Git client written in Rust. An in-progress side project to go deeper on systems programming and ergonomic TUIs.
I'm Nicolás, a backend engineer who likes the problems that show up when things fail: distributed systems, idempotency, integrations that have to recover on their own. I work mostly in TypeScript and NestJS on GCP and Kubernetes, with Python and Rust for fun. Outside the day job I ship side projects (AI agents, data platforms, mobile apps) to keep learning across the stack.
Education
2019 — 2023
B.S. Software Engineering (Ingeniero en Tecnologías Computacionales)
Tecnológico de Monterrey
Campus Monterrey · focus areas include Information Security
2016 — 2019
High School Diploma
Tecnológico de Monterrey
Languages
SpanishNative
EnglishC1 / Advanced
FrenchBasic
Outside work
Crypto trading & market analysis
Building & upgrading PC hardware
Automating everything
Video games
STEP 07 · Contact
Let's build something that holds up when things fail.
Open to hybrid and remote roles. The fastest way to reach me is email or LinkedIn.
