Job Title

Company DescriptionMoleculyst Ventures Private Limited is a deep-tech biotechnology startup operating at the intersection of computational biology, artificial intelligence, and genomics, with a mission to enable precision and personalized healthcare.We focus on modeling complex biological systems by uncovering interactions between genetics, molecular structure, and environmental factors. Our work spans AI-driven diagnostics, therapeutic discovery, and preventive healthcare, leveraging graph-based representations, large-scale biological datasets, and first-principles reasoning. Moleculyst aims to empower clinicians and researchers with scientifically grounded, next-generation computational tools.Role Description – Deep Learning Specialist (GNN-Focused)This is a remote contract role for a Deep Learning Specialist with prior hands-on experience in Graph Neural Networks (GNNs) and a strong mathematical foundation.The role involves designing, implementing, and optimizing deep learning models for large-scale biological data, including molecular, genomic, and interaction networks. You will work closely with interdisciplinary teams spanning biology, chemistry, and data science to develop theoretically sound and practically effective models for real-world biomedical problems.This is not a plug-and-play ML role — we are looking for individuals who understand why models work, not just how to run them.Key ResponsibilitiesDesign and implement deep learning architectures, with a strong emphasis on Graph Neural Networks (GNNs) for biological and molecular dataDevelop models for molecular graphs, protein/peptide structures, interaction networks, and omics dataResearch, evaluate, and implement state-of-the-art methods (message passing, equivariant models, attention mechanisms, representation learning)Perform rigorous model evaluation, ablation studies, and error analysis grounded in both math and domain contextAutomate training, evaluation, and experimentation pipelines for large datasetsCollaborate with biologists and domain experts to translate biological hypotheses into computational modelsAnalyze results to extract scientifically interpretable and actionable insightsRequired Qualifications (Non-Negotiable)Deep Learning & MathematicsStrong understanding of deep learning fundamentals, including:Linear algebra, probability, optimizationBackpropagation, loss functions, regularizationBias–variance tradeoffs and generalizationProven experience with Graph Neural Networks, such as:GCN, GAT, MPNN, Graph Transformers, or related architecturesPractical understanding of message passing, aggregation, and graph representationsAbility to derive, reason about, and modify model formulations, not just apply librariesProgramming & FrameworksStrong proficiency in PythonHands-on experience with PyTorch (preferred) and/or TensorFlowFamiliarity with ML/DL tooling for experimentation, logging, and reproducibilityBiological Data ExperiencePrior experience handling biological or molecular datasets (e.g., genomics, proteomics, peptides, molecular graphs, interaction networks)Familiarity with bioinformatics workflows, data preprocessing, and domain constraintsAbility to bridge biological intuition and mathematical modelingPreferred / Advantageous QualificationsExperience with equivariant or physics-informed models (e.g., geometric deep learning)Exposure to protein language models, molecular descriptors, or structural biology dataExperience deploying or training models on cloud platforms (AWS, GCP, Azure)Master’s or Ph.D. in Computer Science, Bioinformatics, Computational Biology, Applied Mathematics, Physics, or related fieldsWhat We Value MostDepth over buzzwordsMathematical clarity over black-box usageCuriosity, rigor, and the ability to question assumptionsPrior real work on GNNs, not just coursework or tutorialsIf someone cannot explain their model mathematically, they are not a fit.