About us
At Diabolocom, we build AI systems that operate on real-world customer conversations across voice and text channels. These interactions are complex, highly unstructured, and require systems that function effectively even with limited labeled data.
As an AI Research Engineer, you will drive our scientific roadmap. You will be responsible for investigating, prototyping, and validating novel approaches to NLP and Agentic AI. Your primary focus will be tackling Diabolocom’s open-ended challenges including knowledge distillation, small LLM architectures, robustness to noise and artifacts (coming from ASR transcriptions for instance), few-shot entity extraction, reliable agentic planning, multimodal LLMs—translating state-of-the-art papers into applied solutions. You will bridge the gap between theoretical research and product value, while actively contributing to the wider scientific community through publications and open-source projects.
As part of our team, you will:
Solve "Unsolved" Problems: Investigate and resolve complex challenges where no standard solution exists, specifically focusing on noise-robust NLP and data-efficient learning.
Innovate in Data Synthesis: Design and manage methodologies for data generation, building pipelines to synthesize and process datasets for training and evaluation where real-world data is scarce or unlabeled.
Advance Agentic Architectures: Research and implement grounded agent systems, exploring techniques like ReAct, Chain-of-Thought, and tool-use optimization to reduce hallucination and improve planning reliability.
Define Success: Design scientific evaluation protocols and benchmarks that go beyond standard metrics (like accuracy) to measure real world performance.
Drive Knowledge: Stay current with the state-of-the-art in NLP and Agentic AI, contributing to internal knowledge sharing and external publications such as blogs or scientific articles.
We’ll be happy to bring you on board if you have:
A track record of tackling open-ended ML problems, with the ability to navigate ambiguity and design experiments that validate hypotheses.
Excellent proficiency in Python, with a history of writing clean, maintainable, and modular code.
Strong familiarity with deep learning frameworks such as PyTorch or TensorFlow (fluency in both is a plus), alongside the modern NLP ecosystem.
Our ideal candidate would also have experience with:
Data-Centric AI: Techniques like Active Learning, Weak Supervision, or Synthetic Data Generation. You understand that data curation is a research activity, not just a maintenance task.
Advanced Tuning: Instruction Tuning, RLHF/DPO (Direct Preference Optimization), or Parameter-Efficient Fine-Tuning (LoRA/QLoRA).
Agentic Frameworks: Patterns and tools such as LangChain, LangGraph, BAML, or custom tool-use implementations.
Speech Processing: Since part of the data we work with come from noisy speech transcriptions, exposure to ASR (Automatic Speech Recognition) systems or research experience in robust NLP for noisy/spoken text.
Thought Leadership: A history of published papers, impactful technical blog posts, or novel open-source projects.
What we offer:
Research with a Landing Zone: Bridge the gap between theoretical breakthroughs and production-grade reality, solving high-stakes problems at scale.
Contribute to the Frontier: Stay at the cutting edge of the field by contributing to and publishing state-of-the-art (SOTA) research.
An AI-Native Environment: Join a company where AI is the fundamental engine of our strategy, not a peripheral experiment or a "bolt-on" feature.
High-Caliber Collaboration: Work alongside a veteran team of researchers and engineers who value rigorous technical standards and radical autonomy.
Work Where You’re Best: We offer a "results-only" culture with flexible arrangements and a remote-first mindset.
Flexible working arrangements and remote work options.
Recruitment Process
1.Introductory call with a Talent Acquisition Manager
2.Take-home assignment (48-hour window)
3.Final interview with Kevin, Head of AI R&D at Diabolocom
Please ensure that you complete the questionnaire in full when submitting your application. Applications without completed questionnaire responses will not be reviewed.
We look forward to discovering your work.
