↓ Resume (PDF)   ·   ↓ Long CV (PDF)

Current role

2025–now

Senior Research Scientist · Agigo AG

Zürich, Switzerland

Production speech-and-audio LLMs, synthetic conversational data, GPU-efficient multi-client inference.

Experience

2024–2025

Research Engineer · Telepathy Labs

Zürich, Switzerland

Speech recognition, understanding, and generation for conversational AI agents.

Summer 2023

ML Engineer Intern · Apple

Cambridge, MA

Discriminative training of language models for ASR on tail named-entity data.

Spring 2023

Applied Scientist Intern · Amazon Web Services

Seattle, WA

Joint speech-to-text translation and transcription research. Work published at EMNLP 2023.

2020–2024

PhD Researcher · Idiap Research Institute & EPFL

Martigny, Switzerland

Thesis: "Low-Resource Speech Recognition and Understanding for Challenging Applications." Advised by Dr. Petr Motlicek and Prof. Hervé Bourlard.

2019–2020

Research Engineer · Idiap Research Institute

Martigny, Switzerland

ATCO2 project (SESAR JU · EU Horizon 2020). Automatic speech recognition and contextual understanding for air traffic control.

Education

2024

PhD · EPFL & IDIAP

Lausanne / Martigny, Switzerland

Computer Science. Dissertation on Automatic Speech Recognition for Air Traffic Control, domain shift, self-supervised pretraining, and contextual biasing.

2017–2019

MSc · Erasmus Mundus EU4M

Oviedo, Spain · Nancy, France · Cluj-Napoca, Romania

Mechatronics & Micro-Mechatronics. Thesis on computer vision for breast cancer diagnosis (SBRA EU project, Universidad de Oviedo).

2011–2016

BSc · Universidad Autónoma del Caribe

Barranquilla, Colombia

Mechatronics Engineering.

Awards

  • Best Student Paper nominee: Interspeech 2023 (CommonAccent)
  • 1st place: International Create Challenge 2020 · HealthTech Award (Groupe Mutuel)
  • Erasmus Mundus Scholarship: EU Commission (EU4M programme, 2017)
  • DAAD Research Scholarship: Germany (2014)

Skills

Core · Python · PyTorch · SpeechBrain · Kaldi · HuggingFace · LaTeX · Git · Linux · GPU training & inference · LLMs · ASR · TTS · Self-supervised learning · NLP

AI-assisted development · I use agentic coding tools, primarily Claude Code, as a core part of my engineering workflow, applying them from low-level GPU kernels up to high-concurrency serving:

  • Kernel development: CUDA Graphs, Triton kernels, and torch.compile for GPU-efficient inference
  • TTS & LLM systems: streaming synthesis, controllability, and omni-modal model serving
  • High-concurrency deployment: throughput and latency optimization for multi-client production serving
  • Rapid prototyping: from idea to working proof of concept in hours rather than days
  • Data curation: large-scale filtering, cleaning, and synthetic data generation pipelines
  • Evaluation & infrastructure: automated benchmarking harnesses and reproducible experiment tooling

Languages

Spanish (native) · English (fluent) · French (intermediate)

Contact

juan.zuluaga@eu4m.eu · GitHub ↗ · Scholar ↗ · LinkedIn ↗