Attune Most AI systems today do not understand people.
They predict language patterns without truly reasoning about human behavior.
Attune exists to change that.
We are building an AI system that learns from real conversations over time, identifying patterns in how people communicate, collaborate, and make decisions.
The goal is to turn raw interaction data into structured understanding and practical, context-aware coaching.
This is not a prompt layer or a wrapper.
The system combines longitudinal memory, behavioral modeling, uncertainty-aware reasoning, and adaptive interaction.
The product is already live.
The focus now is on making it meaningfully more intelligent, reliable, and grounded in how humans actually behave.
About Unlimit Blue Health is incubated by Unlimit (https://unlimit.
Ventures/), an innovation consultancy firm dedicated to accelerating the innovation roadmap of large companies through startup/lean methodologies.
The team includes ex-Mc Kinsey consultants, experienced startup founders (with products reaching 2 M+ active users), and innovation leaders.
In Portugal, Unlimit works with some of the most innovative companies (NOS, EDP, and others), and globally collaborates with startups across AI, Fintech, Cleantech, and related industries.
Who You Will Work With You will work directly with: leading researchers from FEUP AI Lab operators and builders working at the intersection of AI and psychology founders with experience building a Y Combinator-backed company and 2 M+ active users The Role This is a founding-level research role focused on building an intelligence system that operates in real-world conditions.
You will work across applied research and system design, defining how the system models human interaction, tracks signals over time, and reasons under uncertainty.
Your responsibility is to bridge research and product.
Not just to explore ideas, but to make them work in practice.
What You Will Do Design and evolve the reasoning and memory systems behind Attune Develop models that capture structure in human communication and behavior Build approaches for longitudinal signal tracking across conversations Improve how the system handles ambiguity, uncertainty, and incomplete data Define evaluation frameworks that go beyond standard benchmarks Prototype and test new methods across LLM and non-LLM approaches Contribute to core architectural decisions across retrieval, memory, and inference How You Work You approach problems from first principles and are comfortable questioning assumptions You can translate research ideas into working systems You are comfortable operating without clean datasets or predefined problem statements You balance rigor with pragmatism and care about how systems behave in real use You communicate clearly and can collaborate across technical and non-technical domains This Role Is Likely a Good Fit If You have Phd education on Computer Science, AI, Machine Learning, NLP, or related fields You have built or designed AI systems beyond simple API integrations You are interested in problems at the intersection of human behavior, reasoning, and machine intelligence You enjoy working on open-ended problems where the solution space is not predefined You want to build systems that operate in the real world, not just on benchmarks Nice to Have Experience with longitudinal modeling or behavioral data Background in computational linguistics, cognitive science, or human-computer interaction Experience with retrieval systems, memory architectures, or knowledge graphs Experience working on early-stage AI products Why This Is a Rare Opportunity You will define the intelligence architecture of the product from an early stage You will work on real human interaction data rather than synthetic tasks You will operate at the intersection of research and product, with direct feedback loops You will have high ownership and direct impact on both system design and company direction This role is not for someone looking for well-scoped problems, clean datasets, or incremental improvements.
