Skip to main content
Andrés Muelas TenorioAT
Looking for a job

Andrés Muelas Tenorio

@andresmuelastenorio

AI Engineer | Python, LLM agents, RAG and MCP | LangGraph, FastAPI, AWS | Creator of CoKarTa (App Store)

Spain
Message

What I'm looking for

Junior AI Engineer, Machine Learning Engineer, Backend Engineer (Python) or Software Engineer role. Entry-level, remote, from Spain or Europe. Open to full-time, part-time, contract or internship. Available immediately.

AI Engineer based in Logroño, Spain. I build LLM applications in Python and like to take them end to end, from the first version all the way to production.

I've always liked computers, and science and maths led me to study Industrial Electronics and Automation Engineering at the University of La Rioja. In late 2024, a few months before my first internship, I started programming. My older brother works in software, and through him I got to know the field. When I discovered generative AI, everything took off, and that's when I fell in love with software.

Since then I've kept learning on my own, mostly by shipping projects. I've been building with Claude Code since February 2026 and have studied its architecture in depth (subagents, skills, hooks and loops). Alongside that, I'm preparing for the Cambridge C1 Advanced.

Projects

BackedCV: AI job-fit checker with verified CV edits (Oct 2026 – Present)

Built solo during GREVIX AfterCode 2026: it says whether to apply to a job, what could screen you out and how to tailor the CV without inventing experience. Code checks and an independent LLM verifier block 20 of 20 invented rewrites and pass 18 of 20 faithful ones.

CoKarTa: AI nutrition app (Jul 2026 – Present)

Built and launched an AI nutrition app that logs meals from a sentence or a photo. On the App Store, with 15 users currently on the free version.

Built a streaming agent with 17 tools. Batched writes cut the calls to log six meals from 26 to 6. Cut write-turn latency by 60% by lowering the model's reasoning effort; prompt caching serves 98.8% of the prompt.

Built an MCP server with 18 tools and its own OAuth 2.1 server, so the app works from inside Claude or ChatGPT. Added guardrails that got the app through Apple's App Review after an initial rejection, plus per-user AI spend control. In production on AWS with Docker and GitHub Actions CI/CD, with OpenTelemetry tracing to Langfuse and 2,572 tests.

AMG: eval harness for AI agents (2026 – Present)

Built the evaluation harness for the CoKarTa agent: real model calls, blind LLM judges and an optimizer that keeps a change only if the score improves. Raised the pass rate from 44 to 49 of its 49 original scenarios in 11 cycles (9 changes kept, 2 reverted); the harness also caught a change that broke 2 passing scenarios. Judge calibrated against 8 known verdicts (8 of 8); a run of 68 scenarios costs $0.08. MIT-licensed.

MISSELA: multi-account email client (Nov 2025 – Aug 2026)

Gmail and Outlook: FastAPI and PostgreSQL backend (48 migrations), PostgreSQL-backed job worker, OAuth 2.0, React and TypeScript frontend, 3,970 tests.

Honors and awards

Adaptive Legend – GREVIX AfterCode 2026 (Oct 2026)

1 of 2 special awards among 17 prizes at an online hackathon from a university in India with 1,200+ registered participants, for the best adaptation of BackedCV to the event's challenges.

Experience

Work history, roles, and key accomplishments

NTT DATA logoND

Generative AI Engineer Intern

Apr 2026 - Oct 2026 (6 months)

Part of a GenAI team that built a claims assistant for an insurance client.
Contributed to its LangChain RAG pipeline: chunking, embeddings, pgvector, hybrid search, reranking.
Took part in building its LangGraph agent with tool calling and MCP server.
Helped with evals (LLM-as-a-judge, LangSmith) and prompt/retrieval tuning.
Took part in deploying it on AWS with FastAPI, Docker and GitLab CI/CD.

PanoImagen logoPA

Backend Developer Intern

PanoImagen

Jun 2025 - Sep 2025 (3 months)

Sole developer of a Python computer vision system still in production at a factory: an Intel RealSense depth camera over a conveyor belt turned into a line scanner.
Stitched 100 px strips using frame timestamps and belt speed, detected parts by height and measured them in millimetres after calibrating the camera (0.46 px error), with OpenCV, NumPy, pyrealsense2, threading and ThreadPoolExecutor.

Education

Degrees, certifications, and relevant coursework

University of La Rioja logoUR

University of La Rioja

Bachelor's degree, Industrial Electronics and Automation Engineering

2022 - 2027

Languages: Spanish (native); English (B2 upper-intermediate, preparing for Cambridge C1 Advanced).

Get matched with your dream remote job

Sign up now and join over 1,000,000+ remote workers who receive personalized job alerts, curated job matches, and more for free!

Sign up
Himalayas profile for an example user named Frankie Sullivan