Feathersoft
AI/ML Platform Architect
Job Description
- Help architect, design and implement ThinkBio.Ai AI insight and data platforms; design agentic architecture for the AI platform
- Evaluate and analyze foundation models for protein structure, clinical data, genomic data and small molecule data; fine tune foundation models. Design loss functions relevant to biological applications
- Design scalable data models for multi-modal biological data
- Create a componentized architecture with component level structured APIs
- Design and implement knowledge graphs for biological data; apply deep learning based methods to create representations from knowledge graphs
- Work closely with other team members and partners to identify most critical data centered challenges and address them using cutting-edge computational, statistical and machine learning applications
Requirements
- Ph.D. in Computer science, AI/ML, Mathematics/Statistics
- 10+ years’ experience and technical expertise in AI platforms, data platforms, deep learning models
- Solid understanding of computer science fundamentals
- data structures: trees, directed and undirected graphs, hash tables, heaps
- algorithms: search, sort, rank, graph based algorithms
- Solid understanding of statistical/mathematical concepts
- probability theory
- correlation metrics
- Bayesian probability
- Statistical regression methods
- Good understanding of GPU architectures – CUDA
- Proficiency in Python, C, C++
- Expertise in AI frameworks: PyTorch, TensorFlow
- Exposure to agentic frameworks like LangChain and LlamaIndex
- Firm grasp of modern statistical methods and machine learning techniques, and their applications to large-scale, data
- Ability to manage projects with minimal supervision, using creative and analytical thinking.
- Ability to drive highly collaborative work across the organization and outside the company
- Excellent oral and written communication skills
Preferred Skills
- Experience and understanding of how bioinformatics and data science can best be applied to speed up drug discovery
- Basic understanding of biological concepts and a familiarity with drug development process
- Knowledge of bioinformatic tools and databases to analyze genomics and proteomics data.