About the role
We are looking for a Head of Formal Methods in AI to lead our advanced AI initiatives. This role demands a Masters / PhD in Mathematics or Computer Science, with experience in design and implementation of neural networks, beyond finetuning/adapting LLMs. The candidate will have used different machine learning architectures models to solve different real-world and hard engineering problems, developed models for various problem areas (e.g. vision, language, graphs), and also trained them with various techniques (e.g. value models, reinforcement learning, search strategies). Furthermore, a robust understanding of dependent type theory and formal logic, along with proficiency in functional programming languages will be incredibly valuable to guide research and development efforts. Reporting to the VP of AI, the Head of AI Research will spearhead theoretical advancements and translate them into practical applications.
Key responsibilities
- Lead groundbreaking AI research focusing on neural networks, machine learning algorithms, dependent type theory, and formal logic.
- Develop fresh, practical, scalable approaches to AI system design to improve reasoning, accuracy, and explainability/interpretability.
- Explore ways to train neural networks (NN) to explore/traverse knowledge graphs in generating predictions.
- Develop practical solutions to tackle a domain’s knowledge representation, invariance, and other problems, and tailor data sets to enable efficient machine (neural network) learning and output generation.
- Apply various ML techniques, including supervised, unsupervised, and reinforcement learning to drive model performance, and tweak commonly used methods to fit our needs based on deep theoretical understanding and empirical experience.
- Optimize compute usage for model training/test by maximizing output from smaller compute-efficient models before moving to larger ones.
- Establish and nurture partnerships with academic institutions, research entities, and industry peers to promote collaboration and knowledge sharing.
- Publish and present research findings in prestigious journals and at industry conferences, enhancing our reputation and influence in the AI field.
Desired Experience & Qualifications
- Masters / PhD in Computer Science, Mathematics, or a related field with a strong emphasis on AI, machine learning, or theoretical computer science.
- Demonstrated success in AI research, including the design and deployment of neural networks and machine learning models, especially deep understanding of different neural network (NN) architectures, their pros/cons, and applicability to different problem types.
- Hands-on direct experience of using, training, reinforcement learning, and optimizing transformer models of varying sizes and designs.
- Proficiency in programming languages and tools relevant to AI research, such as Python, TensorFlow, or PyTorch.
- Solid understanding of dependent type theory and formal logic, with practical applications in AI and computer science.
- Familiarity with formal functional programming languages, verification frameworks (e.g. Haskell, Prolog, Z3, LEAN), with experience applying these in research or practical projects.
- Familiarity with formal verification, symbolic computation, or related areas.
- A strong publication record in top-tier AI and computer science journals or conferences.
Hummingbird Bioscience is firmly committed to Equal Employment Opportunity (EEO) and to compliance with all Singapore and US federal, state and local laws and/or guidelines that prohibit employment discrimination on the basis of age, race, color, gender, sexual orientation, gender identity, ethnicity, national origin, citizenship, religion, genetic carrier status, disability, pregnancy, childbirth or related medical conditions, marital status, protected veteran status and other protected classifications.
