Machine Learning Research Scientist — Generative Antibody Design

Help build the next generation of AI for antibody discovery. mAIbe is advancing machine-learning methods at the interface of protein science and therapeutics. If you are passionate about generative AI, drug discovery, structural biology, computational immunology and chemistry, this is your chance to work at the very frontier. Role We're hiring an ML Research Scientist to push the frontier of generative antibody design. You will join a team of ML scientists and engineers from startups, industry, and academia to explore and extend SOTA architectures—including sequence–structure co-diffusion, flow-matching for flexible antigens, AI-augmented docking, antibody PLMs, paratope/epitope inference, affinity prediction, and humanization/immunogenicity modeling. This role combines deep theoretical understanding with hands-on experimentation. You will design and prototype new algorithms, run careful experiments, and translate promising ideas into validated methods that advance our discovery pipeline. Partnering closely with protein engineers and immunologists, you'll ensure model outputs are biologically interpretable and experimentally meaningful. What you'll do Research & prototype new generative models for antibodies (e.g., sequence–structure co-diffusion, flow-matching with target conditioning, docking-aware denoising). Build antibody PLMs and representation-learning methods specialized to CDRs; integrate them as guidance signals for generation. Model binding and function: develop affinity predictors from sequence and structure; integrate physics- or ML-based scoring into training/sampling loops. Docking & complexes: combine AI with physics docking and learned rescoring for robust Ab–Ag pose modeling. Safety & developability: create immunogenicity, humanization, and multi-objective optimization modules to balance affinity with manufacturability. Own the loop: design datasets/splits, run large-scale training (cloud/HPC), and partner with experimental teams for prospective validation. Publish & open source: write papers, release code/models, and present at top ML/CB venues. Minimum qualifications PhD (or equivalent research experience) in ML/CS/Applied Math/Comp Bio/Physics, with peer-reviewed publications or strong open-source track record. Deep expertise in generative modeling (diffusion or flow-matching) and/or equivariant architectures (SE(3), E(n) GNNs) or similarly impactful paradigms. Strong Python + PyTorch/JAX engineering; comfort with large-scale training, profiling, and distributed compute. Solid grounding in at least one of: protein structure prediction, docking/scoring, molecular simulation, or biological sequence modeling. Bonus points (nice-to-haves) Antibody-specific know-how: CDR definitions (Kabat/IMGT), numbering (ANARCI), SAbDab/OAS data, repertoire statistics. Experience with AlphaFold-Multimer/AF3-class models, Rosetta/PyRosetta, OpenMM, or docking tools (e.g., ZDOCK, ProPOSE). Graph/geometry tooling (PyTorch Geometric, e3nn), and experience with multi-objective optimization for design. MLOps: Docker, Weights & Biases, SLURM/K8s, cloud (GCP/AWS/Azure). Cross-functional collaboration with wet-lab teams (assays, BLI/SPR, NGS libraries). What we offer Impactful problems: your models will impact real-world health challenges and directly influence antibody and vaccine design campaigns. Serious compute & data: access to curated antibody/antigen datasets and modern training infrastructure. Scholarly support: backing for conferences, publications, and patents. Publication & OSS support: time and resources to publish and release tools. Growth & ownership: lead projects end-to-end, mentor interns/engineers, shape our roadmap. Competitive compensation and benefits. We promise to Communicate clearly at every stage. Focus on what you can grow into—not just what you've done before. Be transparent with feedback and open to yours. Remote or On-Site Biotech meets AI. #J-18808-Ljbffr