ML Researcher

About Virtido

Virtido is an entrepreneurial and innovative IT company headquartered in Zurich, Switzerland. We realize ideas and projects - from strategic concept to technical implementation closely alongside our dynamic clients with a strong focus on start-up or fast-growing companies. Since inception in 2015, we have grown rapidly to currently 140+ professionals in Switzerland, Poland and Ukraine.

About Our Client

Our client is a sales and marketing AI platform designed to help growing businesses scale their sales and marketing efforts. Their platform enables users to hire, train, manage, and scale an AI sales team in minutes instead of weeks. With a mission to make sales superpowers accessible to everyone, they provide tools that streamline the sales process and enhance customer engagement.

About This Role

Were looking for a Machine Learning Researcher with strong experience in large language models (LLMs) and applied machine learning. You'll work on building and refining cutting-edge models across text, speech, and multimodal domains, contributing to products and research that push the boundaries of whats possible with AI. While speech recognition is one of the domains we work in, the primary focus is on language understanding, model alignment, and generative model development.

Key Responsibilities

• Conduct research in natural language processing, generative AI, and large language models, including fine-tuning and task-specific adaptation.
• Develop and experiment with models for tasks such as text summarization, semantic search, dialogue systems, information extraction, and multimodal understanding.
• Collaborate with teams working in speech recognition, contributing expertise where text and speech intersect (e.g., transcribed dialogue, audio-to-text pipelines).
• Design and evaluate model training strategies, including supervised, unsupervised, and reinforcement learning approaches.
• Contribute to deployment-ready research and help transition models from prototype to production.
• Analyze experimental results and optimize for both accuracy and efficiency across different modalities.

Requirements

• Solid understanding of deep learning, transformers, and modern NLP techniques.
• Hands-on experience with LLM training/fine-tuning, using frameworks like Hugging Face Transformers, OpenAI API, or custom pipelines.
• Strong Python skills and experience with ML libraries such as PyTorch, TensorFlow, or JAX.
• Familiarity with training and evaluating models on large-scale datasets.
• Good level of English language and strong communication skills - you will be working in an English speaking team. 

Nice to Have

• Experience with speech recognition frameworks (e.g., Whisper, wav2vec, DeepSpeech) or integrating ASR into NLP pipelines.
• Background in multimodal models (e.g., audio + text, vision + language).
• Knowledge of RLHF, prompt engineering, or instruction-tuning for LLMs.
• Contributions to academic publications or open-source projects in ML or NLP.

Why Join

• Work on impactful AI problems at the frontier of research and application.
• Collaborate with a high-caliber, interdisciplinary team in a fast pace start-up environment.
• Ability to work fully remote or from one of Virtido offices 
• Courses of English and German
• Flexible schedule options.