# Diego Valencia Román

> Software Engineer · Applied AI · Distributed Systems. I build real-time distributed systems and take AI from prototype to production, designing the architecture and writing the code.

## About
I'm a software engineer with 7+ years building backend platforms, cloud architectures and machine-learning products. At Microsoft I work on distributed systems for real-time data that power AI agents and live communications.

Before that I co-led engineering at WIND Garage, COPEC's innovation hub, where we built a reusable backend framework that let new businesses launch MVPs 40% faster and delivered software, AI and infrastructure for 30+ ventures. At Barracks.gg I shipped real-time data products and ML models seen by millions of Riot Games esports viewers.

I like turning fuzzy ideas into systems that scale: choosing the architecture, building it, and helping teams adopt AI pragmatically. I've also taught Machine Learning, Deep Learning, Big Data and Software Engineering at university level.

## Experience
- **Microsoft**, Software Engineer II (Aug 2024 – Present)
  - Build distributed systems for real-time data and metadata that enable AI agent and live communication scenarios.
  - Develop Java and Kafka Streams services for session-state management and event-driven coordination, including state expiration and command delivery to downstream services.
  - Implement .NET request routing and gRPC service communication with freshness checks and event-based fallback paths.
- **WIND Garage · COPEC**, Software Engineer (Jul 2022 – Aug 2024)
  - Co-created a backend framework that let new businesses launch MVPs 40% faster.
  - Co-led engineering teams delivering software, AI and infrastructure for 30+ new businesses across mobility, convenience and energy.
  - Helped small and medium businesses adopt and implement AI tools.
- **Barracks.gg**, Software Engineer · Machine Learning Engineer (Mar 2021 – Jul 2022)
  - Delivered end-to-end products bringing real-time gameplay data to millions of viewers of Riot Games esports livestreams; deployed live fixes at international onsite events.
  - Led a project that cut local service latency by 200 ms and built tools to monitor data-to-screen time.
  - Built livestream detection and classification models and a brand-tracking web app, deployed with PyTorch, GCP AutoML, Vision AI and Kubernetes.
- **Independent consulting**, Freelance Software Engineer (Jun 2019 – Feb 2021)
  - Lüqfit: moved a health services company from offline to remote care with a MERN app for video consultations, booking, messaging and payments, deployed on GCP with Jenkins.
  - Mediaweb: automated streaming-service provisioning with React, Node.js and MongoDB.
  - University health department: built student appointment booking with PHP, Oracle and JavaScript.

## Selected work
- **Bloom Academy** (Independent AI project · 2025): An AI learning app prototype. Uploaded PDFs are ingested, embedded with OpenAI and indexed in Pinecone for retrieval-augmented answers via LangChain, plus a realtime voice tutor built on the OpenAI Realtime API over WebRTC.
- **Real-time esports data on air** (Barracks.gg for Riot Games · 2021–2022): Live gameplay data on screen for millions of esports viewers. Cut local service latency by 200 ms at international events and built monitoring for data-to-screen time.
- **Livestream brand tracking** (Barracks.gg · 2021–2022): Computer-vision models that detect and classify sponsor exposure in livestreams, with a web app that replaced manual data collection for clients.
- **MVP backend framework** (WIND Garage · COPEC · 2022–2024): A reusable backend framework that let new ventures launch MVPs 40% faster, used across mobility, convenience and energy businesses.
- **Telehealth platform** (Lüqfit · 2019–2021): Took a health services company from offline to remote care: video consultations, booking, messaging and payments.

## Teaching
- **DUOC UC**, Part-time Professor (Mar 2023 – Aug 2024): Big Data, Software Engineering, Deep Learning and Machine Learning.
- **Universidad de Tarapacá**, Part-time Professor (Mar 2022 – Jan 2023): Programming II in Python and C for computer science and data science programs.

## Education
- **Universidad de Chile**: Postgraduate Certificate in Artificial Intelligence (2021 – 2022)
- **Universidad de Tarapacá**: Ingeniería Civil en Computación e Informática (Computer Engineering) (2014 – 2021)

## Skills
- Languages: Python, TypeScript, JavaScript, Java, C#, Go, SQL
- Distributed systems & cloud: Kafka / Kafka Streams, gRPC, Kubernetes, Docker, Google Cloud (GCP), Redis, MongoDB, Oracle
- AI & ML: RAG, LangChain, OpenAI API & Realtime, Pinecone, PyTorch, GCP AutoML, Vision AI
- Frameworks: .NET, Node.js, NestJS, Express, React, Angular

## Services
- **Software & cloud architecture**: Event-driven, real-time backends that scale: Kafka, gRPC, Kubernetes and GCP, designed for observability from day one.
- **AI implementation**: RAG, AI agents, realtime voice and computer vision, taken from prototype to production and wired into your existing systems.
- **MVP to scale**: Ship a focused MVP fast, validate it, and keep a clear path to a production-grade platform.
- **Mentoring & training**: Hands-on AI and software engineering workshops for teams, plus one-on-one mentoring.

## Contact
- Website: https://diegovalencia.cl
- Email: dvr.diegoz@gmail.com
- LinkedIn: https://www.linkedin.com/in/diego-valencia-4982a0106/
- WhatsApp: https://diegovalencia.cl/api/whatsapp (opens a chat)
- Location: Chile · working remotely worldwide
- Languages: Spanish (native), English (advanced)