Research Engineer: Graph Machine Learning ...

Research Engineer: Graph Machine Learning Atman Labs,London About Atman Labs At Atman Labs we are building software toemulate proactive human expertise. Emulating human experts withdeep knowledge and proactive assistance has largely been impossibleto do via standalone Artificial Intelligence techniques. As anapplied research and commercialization company we are deploying ourproducts in a number of domains to demonstrate the value of ourapproach – from proactive shopping assistance, to personal teachersto healthcare concierges – and with this commercial focus advanceour unique research that lies at the intersection of ReinforcementLearning rewards, Large Scale Knowledge Representation, andPredictive Models inspired by biological priors. The Next Frontierof Machine Reasoning: Web-scale Knowledge Graph Exploration usingReinforcement Learning Human experts can form and explorestructured mental models in their heads to solve open-endedproblems across different domains. Our research seeks to emulatethis process through a novel combination of using reinforcementlearning agents to perform exploration through a knowledge graph.Knowledge graphs allow us to represent structured information andthe logical relations that govern it, unlocking the ability tobuild reinforcement learning strategies that can learn to solvecomplex, open-ended problems across web-scale andcontinuously-evolving domains. You will be leading the research onknowledge representation and how it can serve to build AI systemscapable of such complex reasoning. You will work on formulatingresearch problems that explore how Reinforcement Learningalgorithms can interact with large and complex knowledge graphs toreason over ambiguous tasks. To do this, you will develop knowledgegraph machine learning techniques that will power several toolswithin our products. Knowledge graph representations (e.g.embeddings) are critical to representing web-scale, structuredinformation in a compact format for a reinforcement learning agent,ensuring scalability. You will lead the efforts on training andvalidating graph embedding algorithms that capture multi-hopsemantics within large web-scale knowledge graphs. Additionally,you will develop link prediction models that will enhance both thereasoning over the knowledge graph and recommendations. About YouWe are looking for ambitious and independent thinkers who have adeep desire to contribute and want to be part of the team thatmakes this a reality for humanity. In order to contribute, youshould have all of these qualities: 1. You have a PhD degree orequivalent industrial expertise in Graph Machine Learning and itsapplications. 2. You have a deep understanding of thestate-of-the-art in graph machine learning, with a focus onlearning graph embeddings and link prediction problems. 3. You haveexperience in training and tuning various graph ML algorithmsincluding GNNs, Message Passing and Graph Transformers. Experiencein building graph-based recommendation systems is a plus. 4. Youhave 5+ years of programming experience in Python and havedevelopment experience with toolkits like PyTorch or Tensorflow andcan deploy models with clean APIs. You are equally capable as asoftware engineer as you are in formulating novel research ideasand your code proves it. Moreover, in order to deeply fit withinour culture, you should embody the following: 1. You are capable ofreasoning from first-principles, where there is no trodden path, aswell as critically evaluate when existing ideas are worthconsidering. 2. You are articulate and can present your ideas inwriting, in person and in small groups educating audiences at alllevels on the application of generative models. 3. You have a high‘faker’ detector in others, and can critically evaluate truth fromfiction in your own work. 4. Your colleagues consider you a highlypositive personality, you amplify the energy of others rather thandampen the mood. 5. Your intensity goes from 0-1000 when you becomeauthentically interested in a topic. 6. You not only have interestsin systems engineering but are deeply curious about a range ofinterdisciplinary topics ranging from computational creativity,knowledge graphs, recommendations, web scale search, deep learning,large language models, computer vision, human consciousness, andthe opportunity to build truly intelligent systems in software thatare inspired by biology. 7. Outside work you can show highcreativity and intensity in your pursuits, you cannot easily becharacterized in one discipline. 8. You consider yourself aninnovator, and original thinker, not a follower. You are lookingfor a way to contribute to the world and want to join our team todo so. You want to work in person in London. We’ll sponsor yourvisa. We have the ambition to usher the world towards co-existingalongside Benevolent AGI. Not only do we believe that our work is acredible approach to functionally emulate human reasoning but webelieve that this mission can also allow us to conceive manycommercial products that yield billions of dollars of commercialrevenues that can support an ambitious R&D effort for years tocome. We are building for a future where humans coexist alongsidebenevolent expert systems and seek to advance the field from thefront. We are looking for ambitious and independent thinkers whohave a deep desire to contribute and want to be part of the teamthat makes this a reality for humanity. Apply with a short messageand a list of your projects, your life story in 5 sentences, yourfavorite book or artist, and your resume to .#J-18808-Ljbffr