Website:
anubio.ai
Job details:
Machine Learning Scientist
Carot labs (dba AnuBio)
đ Chennai/Bengaluru, India
đ https://anubio.ai
đ Full-time
About Anubio:
AnuBio is an AI-driven deeptech company building next-generation multicellular foundation models for immune systems. Our platform, TRAILBLAZERâ˘, creates predictive âdigital twinsâ of patient immunityâenabling discovery of new therapies without predefined molecular targets.
About the role:
Weâre building generative models that predict how cell populations respond to therapeutic interventions, at single-cell resolution and patient scale. Our work sits at the intersection of representation learning, generative modeling, and translational biology, and is used to prioritize treatments and simulate virtual clinical trials. Weâre looking for an ML scientist who can own model architecture and training end-to-end and push the research forward alongside our team. This is a core ML role. Strong modeling instincts and rigor matter most, but curiosity about the biological domain will help you do your best work.
What youâll do:
- Design, implement and train deep generative models (VAEs, transformers, and related architectures) on large-scale, high-dimensional data.
- Shape and constrain latent spaces (e.g. metric learning, contrastive objectives) so that learned latent spaces are controllable and support extrapolation to unseen states.
- Build count-aware decoders and likelihood heads (e.g. negative-binomial / zero-inflated parameterizations) and design composite training objectives with staged optimization.
- Construct careful data pipelines and samplers that prevent shortcut learning and control for confounders.
- Design rigorous benchmarks: zero-/few-shot generalization, ablations, and comparisons against state-of-the-art baselines.
- Translate model outputs into actionable predictions and rankings, and collaborate with domain experts to validate them.
Required qualifications:
- MS or PhD in computer science, machine learning, statistics, physics, or a related quantitative field.
- Strong track record building and training deep neural networks in PyTorch.
- Solid grounding in generative modeling (VAEs, diffusion, flow-based, or auto regressive) and in modern transformer/attention architectures.
- Comfort with the math behind the methods: probability, linear algebra, optimization, and metric/representation learning.
- Experience working with large datasets and the engineering practices (efficient data loading, distributed/GPU training, reproducibility) that make large-scale training tractable.
- Ability to design clean experiments, interpret results honestly, and communicate findings clearly.
Nice to have:
- Experience with single-cell genomics or other high-dimensional biological data such as scRNA-seq.
- Familiarity with computational biology tooling (e.g. Scanpy, AnnData, scVI) and single-cell foundation models.
- Publications or open-source contributions in ML or comp-bio.
What we offer:
Close collaboration with founders and an elite ML team.
Competitive salary.
Apply:
Email your CV and GitHub/Portfolio to smitha@anubio.ai
In the email: "Point us to a model you trained from scratch in PyTorch (not fine-tuned, not a wrapper around a library) and briefly explain one non-obvious design decision you made and the alternative you rejected.â
Anubio is an equal opportunity employer. We hire based on merit and scientific rigor.
Click on Apply to know more.
