Skip to content

Open to AI, Backend & Cloud roles

Rodrigo
Fernandez

AI Engineer | Backend & Cloud-Native Systems

Software Engineer building applied AI systems and the backend infrastructure that runs them in production. I design RAG pipelines end to end — document ingestion, embeddings, vector retrieval and context grounding — and ship them as containerized FastAPI services on Google Cloud Run, working across Python, C# (.NET) and TypeScript.

View selected work Get in touch
Currently
Software Engineer — Backend & AI @ SimiAI
Core Focus
RAG Architectures · FastAPI · Cloud-Native
Based in
Arequipa, Peru — UTC−5
Languages
Spanish (native) · English (B2)
01 Experience

Where I've shipped

Applied AI, backend architecture and cloud-native infrastructure across CAD automation, satellite ML and enterprise services.

  1. SimiAI

    Current

    Remote

    Software Engineer — Backend & AI Automation

    May 2026 — Present
    • Designed and led the proof of concept for a Python agent engine (FastAPI) with tool-orchestration logic, defining the skill registry and data flow for automated engineering design.
    • Built the integration bridge that parses AI-processed field data — photogrammetry, GPS and NLP output — into standardized JSON payloads consumed downstream by the CAD automation pipeline.
    • Programmed C# .NET logic that inserts dynamic blocks (poles, conduits, utility boxes) at precise georeferenced coordinates and orientations derived from routing parameters.
    • Implemented asynchronous WPF/PaletteSet dockable panels (MVVM) so cloud-synced alerts and field media render without blocking the CAD UI thread.
    • C# .NET
    • AutoCAD API
    • Python
    • FastAPI
    • AWS
    • Azure
    • Docker
    • PostgreSQL
    • PostGIS

  2. Committed Organization

    Remote

    Jr. Software Engineer — Cloud & DevOps (CI/CD)

    Nov 2025 — May 2026
    • Deployed and monitored backend inference services across Azure and GCP, building Grafana and Azure Monitor dashboards to track service health and support incident response.
    • Automated cloud infrastructure provisioning with Terraform templates for reproducible, consistent environments.
    • Configured Azure virtual networks, storage accounts and access policies; maintained CI/CD pipelines in GitHub Actions.
    • Azure
    • GCP
    • AWS
    • Terraform
    • Docker
    • Grafana
    • GitHub Actions
    • Linux

    Jr. Software Engineer — Full-Stack

    Apr 2025 — Nov 2025
    • Built and maintained modular backend services in Java and Spring Boot for transactional data flows, containerized with Docker for parity between local and production environments.
    • Java
    • Spring Boot
    • React
    • PostgreSQL
    • Docker
    • Git

  3. Universidad Nacional de San Agustín (UNSA)

    Arequipa, Peru

    Machine Learning Engineer — Research Project

    Jun 2025 — Dec 2025
    • Led development and training of ML models (Python, scikit-learn) for crop classification and phenological stage detection from satellite imagery, reaching +75% accuracy.
    • Built Python/Pandas data pipelines with OpenCV preprocessing for feature extraction, cutting manual analysis time by ~80%.
    • Containerized and deployed the inference services on Google Cloud Run.
    • Python
    • FastAPI
    • Pandas
    • OpenCV
    • scikit-learn
    • Google Cloud Run
    • Docker
02 Selected work

Things I built end to end

Retrieval pipelines, on-device inference and the infrastructure underneath. Every one of these went from empty repo to something someone could actually open.

Law Copilot — RAG over Peruvian Legislation interface

Lead AI & Cloud Engineer · Google AI Partner Catalyst · Jan 2026

Law Copilot — RAG over Peruvian Legislation

Specialized vector knowledge base indexing the Political Constitution of Peru, the Civil Code, the Consumer Protection Code and the Law against Violence towards Women, so answers stay grounded in specific statutes.

Retrieval layer with FAISS and sentence-transformers embeddings + dynamic context injection to cut hallucinations, shipped on Google Cloud Run with multi-stage Docker builds to shrink cold-start latency.

  • Python (FastAPI)
  • LangChain
  • FAISS
  • sentence-transformers
  • Gemini API
  • OpenAI API
  • Next.js
  • Google Cloud Run
  • Docker
Source code for Law Copilot — RAG over Peruvian Legislation on GitHub Live demo for Law Copilot — RAG over Peruvian Legislation
On-Device AI Study Assistant interface

AI Engineer · Google Chrome Built-in AI Challenge · Nov 2025

On-Device AI Study Assistant

4 stars on GitHub

Integrated Gemini Nano inside Chrome to generate summaries, flashcards and five types of exercises fully offline — study documents never leave the user's device.

Deliberate workload split: local Gemini Nano (Prompt API) for privacy-sensitive document processing, and cloud Gemini via FastAPI and LangChain for learning roadmaps and generated games.

  • Next.js 14
  • TypeScript
  • Gemini Nano (Prompt API)
  • Gemini API
  • Python (FastAPI)
  • LangChain
  • PyPDF2
Source code for On-Device AI Study Assistant on GitHub Live demo for On-Device AI Study Assistant
Space Biology Knowledge Engine interface

AI Engineer · NASA Space Apps Challenge · Oct 2025

Space Biology Knowledge Engine

Clustered space-biology literature by meaning using SciBERT embeddings and HDBSCAN, with automatic cluster labeling (KeyBERT, TF-IDF) and an agent that extracts research evidence from each paper.

Exposed the semantic clustering and extracted paper evidence as an interactive knowledge map backed by a FastAPI service.

  • Python (FastAPI)
  • SciBERT
  • HDBSCAN
  • scikit-learn
  • spaCy
  • KeyBERT
  • TF-IDF
  • React
Source code for Space Biology Knowledge Engine on GitHub
DosisFe interface

Public health · Android release

DosisFe

A mobile app for anemia screening that corrects haemoglobin readings for altitude and anthropometry — the adjustment that decides whether a diagnosis in the Andes is right or wrong.

Built with regional health staff as the end users, and shipped as a signed Android release.

  • React Native
  • Expo (EAS)
  • NestJS
  • TypeScript
  • Passport
  • Docker
Source code for DosisFe on GitHub Download APK for DosisFe

More projects

  • AI & Backend Engineer · 2026

    Quechua Q&A — Low-Resource RAG

    Source code for Quechua Q&A — Low-Resource RAG on GitHub

    Question-answering system over Quechua grammar documents, returning answers with the exact source sections and page numbers used so every response is traceable.

    FastAPI backend structured in layered architecture (API / services / repositories / core) with Pydantic models, async I/O, and GitHub Actions CI/CD for both Azure and GCP.

    Python (FastAPI) · Gemini API · Pydantic · Next.js · TypeScript · GitHub Actions · Azure · GCP

  • Projects2Enhance · Team project

    Paqocha

    Private team repository

    A video-based platform for learning Quechua, with topics, captions and translations across Spanish, English and Quechua, plus an admin CMS to manage the whole catalog.

    Clean Architecture on both ends — FastAPI ports/adapters on the backend, a matching adapter layer on the Next.js frontend — deployed on Azure Container Apps.

    Next.js · FastAPI · PostgreSQL · SQLAlchemy · Google OAuth · Azure · Docker

  • Team project · 4 contributors

    TheNextCraft

    Source code for TheNextCraft on GitHub

    An AI product-photography tool — upload a product photo and a background, and Depth Anything V2 estimates the scene depth so the product can be recomposited and re-angled against it, with text layers on top.

    Fully self-hosted infra — Convex, PostgreSQL and MinIO in Docker — with depth estimation running as an async job queue, polled from pending to completed.

    Next.js · Convex · PostgreSQL · Replicate · Depth Anything V2 · Docker · Bun

  • Open source · CLI

    ai-multiworker

    Source code for ai-multiworker on GitHub

    A CLI that offloads token-heavy, low-reasoning work — reading large files, triaging CI logs, checking docs against code — to a cheaper worker model instead of spending an agent's own context on it.

    Three focused tasks (read, log-analysis, docs-audit) after cutting it down from seven — and it reports the tokens it actually saved on every run.

    Node.js · JavaScript · CLI

  • Personal project

    Atlas Sport

    Source code for Atlas Sport on GitHub

    A sportswear e-commerce monorepo — Next.js storefront, NestJS API, PostgreSQL — with a Google-only admin panel and a seeded superadmin bootstrap flow.

    The dev-login shortcut for testing the admin panel is hard-disabled whenever NODE_ENV=production, regardless of any flag left on by mistake.

    Next.js · NestJS · PostgreSQL · TypeScript · Google OAuth · pnpm

  • Team project · 5 contributors

    DeepPeru

    Source code for DeepPeru on GitHub

    A platform for promoting tourism packages across Peru, built as a NestJS + Next.js + Prisma monorepo with a five-person team.

    247 commits across the team under a conventional-commits discipline — the deepest, most collaborative history of anything I've shipped.

    NestJS · Next.js · Prisma · PostgreSQL · TypeScript

Thirty-odd more repositories live on GitHub.

03 Aerospace

A satellite the size
of a soda can

ONISAT CanSat 2026 is a full mission stack I work on end to end: firmware on the flight computer, a LoRa downlink, and the ground software that turns raw packets back into telemetry and stereo images.

The part I'm proudest of: radio packets get lost mid-flight, so the ground station reconstructs the missing image regions with OpenCV inpainting instead of dropping the frame.

Read the source
5 Hz
Telemetry downlink
5
Mission states in the FSM
2
Cameras, red–cyan anaglyph

Signal chain

  1. Onboard

    Sensor array

    Altimeter · IMU · magnetometer · environmental · power

  2. Flight software

    C++17 on Raspberry Pi Zero 2W

    Mission FSM · stereo capture · servo control

  3. Radio link

    LoRa SX1278 → ESP32-S3

    5 Hz telemetry with interleaved image fragments

  4. Ground backend

    Python · OpenCV · WebSockets

    Packet reassembly · inpainting recovery

  5. Dashboard

    Next.js 16 · React 19 · Three.js

    Live 3D attitude · anaglyph stereo imagery

Around the mission

04 Tech stack

What I work with

End-to-end RAG pipelines and ML models on top, Python/FastAPI and C# .NET services underneath, shipped in multi-stage containers across GCP, Azure and AWS.

Highlighted items are what I reach for most weeks — the rest I've shipped with, but less often.

AI & LLM

  • RAG architectures
  • LangChain
  • Gemini API
  • FAISS

Prompt orchestration · Gemini Nano (on-device) · OpenAI API · sentence-transformers

Machine Learning

  • scikit-learn
  • HDBSCAN
  • SciBERT
  • OpenCV

K-Means · spaCy · TF-IDF · Pandas · NumPy

Backend & APIs

  • Python (FastAPI)
  • C# (.NET)
  • TypeScript (NestJS)

Java (Spring Boot) · Flask / Uvicorn · Node.js · REST · WebSockets

Cloud & DevOps

  • Google Cloud Run
  • Docker (multi-stage)
  • Azure
  • AWS (S3)

Terraform · GitHub Actions · Grafana · Azure Monitor · Linux

Databases & Vector

  • PostgreSQL
  • PostGIS
  • FAISS (vector)

MongoDB · SQLite · Prisma · SQLAlchemy

Frontend, Mobile & IoT

  • Next.js
  • React
  • C++17

React Native · Raspberry Pi · ESP32-S3 · LoRa (SX1278) · Three.js

05 Programs

Selected programs & challenges

Where the work was judged by someone other than me.

06 Background

Education & languages

Degree

BSc in Systems Engineering

Universidad Nacional de San Agustín de Arequipa

Arequipa, Peru · Expected 2027

Alongside the degree I run research and competition work through the university: satellite imagery ML for crop monitoring, and the ONISAT CanSat mission.

Languages

  • Spanish Native
  • English B2 — Upper intermediate

Also worth knowing

I work remotely from UTC−5, which overlaps a full working day with both American coasts and a morning with Europe.

Available for new roles

Let's build something
worth shipping.

Applied AI, backend architecture or cloud-native infrastructure — if you're hiring for any of those, or just want to compare notes on RAG systems, my inbox is open. I reply within a day.

rodrygoleu7@gmail.com