Staff AI/ML Scientist (Conversational AI)

We are looking for a highly motivated Staff AI / Machine Learning Scientist with the focus on Conversational AI to join our growing AI Center of Excellence team, responsible for using AI to drive innovation across the organization and accelerate positive impact on our planet and society. In this role, you will redefine how users interact with complex enterprise domain data - transforming simple text interfaces into goal-oriented, multi-step conversational agents for a variety of use cases, from sustainable procurement to supply chain resilience to decarbonisation. You will be part of the team of over 20 scientists and engineers delivering solutions for our product portfolio and enhancing operational excellence.

Your global responsibilities will include (but will not be limited to):

• Design and test multi-turn, state-aware conversational systems using LLM reasoning loops (e.g., ReAct, Plan-and-Solve) and multi-agent orchestration frameworks (LangGraph, AutoGen, or custom state machines)

• Innovate by adapting new frameworks and recent advances in generative and agentic AI

• Promote best practices in AI / ML and mentor junior team members

• Lead functional teams and initiatives to accelerate the adoption of Conversational AI

• Leverage Python, MLflow, Azure stack (e.g., Azure cloud, Azure ML, Azure Data Factory) and Databricks to deliver end-to-end AI solutions to build, benchmark, and maintain AI / ML models

• Stay in close contact with customers to evaluate their needs and recommend solutions that increase customer satisfaction and drive business value

• Stay on top of technology trends in the area of, but not limited to, AI and apply this knowledge to the key challenges ahead of EcoVadis

• PhD or Master’s degree in quantitative discipline (e.g., Statistics/Applied Math, Artificial Intelligence, Computer Science, Computational Linguistics)

• 8+ years of overall experience in machine learning and Natural Language Processing (NLP), with at least 4+ years specializing in Conversational AI in production environments

• Advanced knowledge of agentic frameworks and proven track record building autonomous or semi-autonomous agentic systems at scale

• Experience in leading end-to-end AI initiatives in industry, from design to implementation to integration, and applying best software development practices

• Excellent mentoring skills and accountability for high quality and timely delivery of the initiatives they lead

• Experience of cross-functional collaboration in the development of AI-powered products (e.g., Engineers, Product Managers)

• Experience building conversational interfaces over complex enterprise domains (e.g., Supply Chain, Procurement, Financial Risk) is a plus

• Track record of peer-reviewed publications in top-tier NLP/AI venues (ACL, EMNLP, NeurIPS, AAAI, KDD) is a plus

• Offer available only for candidates eligible to work and live in Spain

• Location: Hybrid in Barcelona (4 days per month in the office) / Full remote from Spain

In return for your expertise, we offer:

• Support with all the necessary office and IT equipment
• Flexible working hours
• Wellness allowance for mental and physical wellbeing
• Access to professional mental health support
• Referral bonus policy
• Learning and development
• Sustainability events and community involvement
• Peer recognition program
• Employee-led resource groups
• Remote work from abroad policy
• Meals and Transportation Vouchers (Coverflex card)
• Dental Benefits
• Life & Accident Insurance + Private Health Insurance
• Paid employee volunteer day
• Paid moving day (1/year)
• Time off: 1 Community Service Day + 1 Personal Day
• Summer Hours in July and August (36 hours per week)
• Hybrid Monthly Allowance for electricity and Internet

Our hiring team looks forward to reviewing your CV, in English, with a guaranteed response to every application. A new job with purpose awaits you!

Don’t fit all the criteria but still think you’d be a good candidate? Please apply anyway to give our hiring team the opportunity to assess your skills and to learn more about what you could bring to EcoVadis. We’re interested in hiring from a diverse and skilled talent pool, regardless of professional and educational background.

Can the hiring process be adjusted to suit my needs? Yes. We want everyone going through the hiring process with EcoVadis to feel confident that you are able to demonstrate your full potential. We welcome applications from disabled people, people with long-term health conditions, and neurodiverse candidates. If you need any adjustments, including the provision of interview questions, please let the hiring team know.

Our team’s strength comes from everyone’s uniqueness and is founded upon mutual respect. EcoVadis commits to equity, inclusion and reducing bias in our hiring processes. EcoVadis does not accept any form of discrimination based on color, national or ethnic origin, ancestry, citizenship, religion, beliefs, age, sex, gender identity, sexual orientation, neurodiversity, disability, parental status, or any other protected characteristic that makes you unique. In your application, we encourage you to remove personal information such as: photographs, marital status, number of children, religion, gender, residential postal code, university graduation date, past medical or parental leave(s) taken, nationality (instead, please state if you are legally eligible to work in the job region/country), university name (instead, please state any degrees obtained and the study major).

Our recruitment processes do not include AI systems that make autonomous decisions impacting individuals’ rights without human oversight. Inputs and outputs of the system have been validated by humans and answers are based on algorithms that analyze previous data with which the system has been trained. We use SmartRecruiters as our applicant tracking system and to schedule interviews. Gemini is used to summarize notes from interviews and to refine job descriptions and social media messaging. AssessFirst is used to support talent assessment decisions.