Skip to content
View vnponce's full-sized avatar
8 followers · 1 following

Achievements

Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: Pull Sharkx3Achievement: Arctic Code Vault Contributor

Achievements

Achievement: Pair Extraordinairex4Achievement: YOLOAchievement: Pull Sharkx3Achievement: Arctic Code Vault Contributor

Block or report vnponce

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
vnponce/README.md

Hi there 👋, I'm Abel Ponce

Senior AI Backend Engineer · Building production AI infrastructure

LinkedIn Email Location Status

🚀 About Me

Senior AI Engineer with 10 years of experience shipping production backend services. Currently at Amazon (AWS — Amazon Connect) building systems that execute automated call flows at high scale, while building a multi-tenant AI Infrastructure Platform on the side — load-tested at 12K+ RPS with p95 <150ms on Kubernetes.

Specialized in AI agents, LLM orchestration, RAG, multi-provider LLM gateways, and event-driven microservices. Background spanning logistics (Nuvocargo), e-commerce (Lovevery), and cloud infra at Amazon. Working remotely from 🇲🇽 Mexico.

  • 🔭 Currently shipping at Amazon Connect Flow — TypeScript · AWS CDK · Lambda
  • 🤖 Building a modular AI Infrastructure Platform — 5 independent services, production-grade
  • 🌱 Going deep on Go, Temporal, Kafka, pgvector, and distributed systems
  • 💬 Ask me about AI agents, RAG pipelines, LLM gateways, or scaling backend services
  • 📫 Reach me: vnponce8@gmail.com

🤖 Currently Building — AI Infrastructure Platform

A modular platform that lets companies integrate production-grade AI without building infra from scratch. Each module is an independent service sharing auth, billing, and observability.

Module What it solves Stack
M1 — AI Gateway Multi-provider LLM routing, cost tracking, no vendor lock-in Go · Envoy · Redis
M2 — RAG Platform Hybrid semantic + BM25 search, 200K+ docs indexed, p95 <280ms Python · FastAPI · pgvector · Kafka
M3 — Agent Orchestrator Durable agent workflows with vector memory, 3M+ executions/day Python · Temporal · pgvector · MCP
M4 — LLM Eval Platform Continuous quality monitoring + drift detection, 2M+ evals/day Python · LLM-as-judge · S3
M5 — Event Mesh Real-time AI inference on event streams, sub-second latency Python · Kafka · Redis Streams

Platform pitch: processing 12K+ RPS with p95 <150ms, validated with k6 load tests on Kubernetes simulating 20K+ concurrent users — using Go, Python, Kafka, Temporal, pgvector, and Envoy.

🛠️ Tech Stack

Languages

Python Go TypeScript JavaScript Ruby

AI / ML

Anthropic OpenAI AWS Bedrock pgvector Temporal MCP

Backend & APIs

FastAPI Node.js Rails Laravel

Cloud & Infrastructure

AWS Kubernetes Docker Envoy Kafka PostgreSQL Redis

Observability & Testing

Prometheus Grafana OpenTelemetry pytest RSpec Cypress

📊 GitHub Stats

GitHub Stats

GitHub Streak

Top Languages

📫 Connect with me

LinkedIn Email GitHub

⭐️ From vnponce — open to freelance & collaboration on AI infrastructure projects

Pinned Loading

  1. healthcare-intake-coordination-system healthcare-intake-coordination-system Public

    A small multi-agent AI system for healthcare intake coordination, built to learn multi-agent orchestration patterns end-to-end without an SDK

    Python 1

  2. football-intelligence-agent football-intelligence-agent Public

    AI-powered serverless agent that answers football questions using AWS Lambda and Claude AI

    JavaScript

  3. mundial-ticket-agent mundial-ticket-agent Public

    small learning project where I explored integrating AWS Bedrock with Lambda to create a virtual assistant for World Cup ticket registrations.

    TypeScript

  4. SeederMX SeederMX Public

    This project shows Laravel model factories to generate 'Estados' 'Municipios' from México.

    PHP 8 9

  5. deliveries-exercise deliveries-exercise Public

    Deliveries exercise with react using TDD

    JavaScript

  6. apartadoenlinea apartadoenlinea Public

    Main idea is to have a sort of ecom but schedulling the orders to specific day, time and store with its owb dashboard to track orders.

    PHP