Senior AI Scientist at regiocom in Lisbon, Lisbon, Portugal

Workster is partnering with a cutting-edge AI startup on a mission to give models the ability to act autonomously—running code, fixing systems, and analysing data—so humans can focus on strategy and creativity. This role is for an experienced AI professional who can take ownership of complex projects, lead the design of advanced features, and mentor others, while driving the evolution of agentic AI systems from research to production.

Your Role

• Lead the research, prototyping, and development of features that advance their Agentic LAM from proof-of-concept to production.
• Drive initiatives in:

1. – Transformers and attention mechanisms.
2. – Fine-tuning and optimisation methods.
3. – Reinforcement learning (especially RLHF and reward modelling).
4. – AI alignment, safety, and interpretability.

• Oversee the full model lifecycle: data ingestion → training → validation → deployment → monitoring.
• Design and implement scalable architectures, including Modular LLMs and Mixture-of-Experts setups.
• Build and maintain evaluation metrics aligned with product and research goals.
• Develop testbeds and behavioural suites to assess long-term agent performance.
• Own the integration of autonomous capabilities such as:

1. – Autonomous code execution.
2. – Structured memory and self-repair loops.
3. – Toolformer-style API integrations.

• Collaborate with engineering teams to ensure robust deployment via Ray, Kubernetes, and CI/CD pipelines.

Your Qualifications

You should bring deep expertise in at least one or two of the following, plus familiarity with the rest:

• Core competencies:

1. – Transformers, attention mechanisms, fine-tuning methods.
2. – Reinforcement learning (RLHF, reward modelling).
3. – Alignment, safety, and interpretability.
4. – Modular LLM design, Mixture-of-Experts, scalable architectures.
5. – Auto-evaluation frameworks.

• Tooling & engineering skills:

1. – Python and PyTorch at an advanced level.
2. – Distributed training, quantised inference, batching, and caching.
3. – Experience with production infrastructure (Ray, Kubernetes, CI/CD for ML).

• Proven ability to lead research initiatives and mentor junior team members.
• Strong track record of delivering AI features from concept to production.
• No strict degree requirements—impact, skills, and leadership are what matter.

The Offer

• Hybrid work in Lisbon with the option for multi-week remote blocks.
• Competitive senior-level salary plus equity potential.
• Subsidised apartment if relocating.
• Health insurance, modern hardware, and a personal learning budget.
• Company-funded travel to global AI conferences and events.
• Significant ownership of strategic AI projects and product direction.
• Work alongside experienced AI engineers and published researchers in a high-impact environment.