Advisor - Antibody Developability Validation & Benchmarking
Eli Lilly and Company
June 25, 2026
On-site
Indianapolis, IN
Clinical Research and Development
Purpose/Role: Validation-led advisor to assess whether TuneLab’s federated antibody models can be trusted to triage real candidates, partnering with antibody modeling scientists on architecture, features, and uncertainty quantification.
Responsibilities:
- Build a canonical antibody developability benchmark suite (aggregation, thermal stability, polyspecificity, self-interaction, viscosity, chemical liabilities, immunogenicity surrogates) and define multi-endpoint rollups to triage decisions.
- Design privacy-preserving, sequence-aware federated test sets with germline/CDR-similarity/clonotype splits; account for shallow vs deep characterization asymmetry.
- Benchmark federated models vs external resources (e.g., SAbDab, OAS, TAP, Jain clinical-stage panel, FLAb).
- Develop cross-domain validation across IgG/bispecific/fragments, expression systems, and assay protocols.
- Implement temporal-split and similarity-aware validation to simulate prospective deployment and detect concept drift.
- Apply statistical rigor (multiple testing, hierarchical structure) with confidence intervals; ensure reproducibility via MLOps.
- Profile performance by germline, CDR length, framework variants, and property ranges; integrate with TuneLab (NVIDIA FLARE).
Basic Qualifications:
- PhD (Computational Biology/Bioinformatics/Chemistry/CS/Stats or related).
- 4+ years post-PhD in antibody discovery/engineering/developability data.
- Experience analyzing antibody developability assays (e.g., HIC, AC-SINS, nanoDSF, polyspecificity, viscosity, chemical liabilities).
- ANARCI (or equivalent) + Kabat/Chothia/IMGT numbering knowledge.
- ML validation protocol design for biological sequence data (similarity-aware splits; held-out tests).
Additional Preferences:
- Protein/antibody language model fine-tuning; sequence liability motifs knowledge.
- Federated learning frameworks (NVIDIA FLARE/Flower/OpenFL/PySyft); uncertainty quantification & calibration; PyTorch/ML ecosystem (HF, scikit-learn, RDKit).
- Antibody structure prediction tools (AlphaFold-Multimer/IgFold/ABodyBuilder); regulatory knowledge; MLflow/W&B; publications and technical writing.
Location/Travel: Indianapolis, San Francisco, or Boston; up to 10% travel.
Responsibilities:
- Build a canonical antibody developability benchmark suite (aggregation, thermal stability, polyspecificity, self-interaction, viscosity, chemical liabilities, immunogenicity surrogates) and define multi-endpoint rollups to triage decisions.
- Design privacy-preserving, sequence-aware federated test sets with germline/CDR-similarity/clonotype splits; account for shallow vs deep characterization asymmetry.
- Benchmark federated models vs external resources (e.g., SAbDab, OAS, TAP, Jain clinical-stage panel, FLAb).
- Develop cross-domain validation across IgG/bispecific/fragments, expression systems, and assay protocols.
- Implement temporal-split and similarity-aware validation to simulate prospective deployment and detect concept drift.
- Apply statistical rigor (multiple testing, hierarchical structure) with confidence intervals; ensure reproducibility via MLOps.
- Profile performance by germline, CDR length, framework variants, and property ranges; integrate with TuneLab (NVIDIA FLARE).
Basic Qualifications:
- PhD (Computational Biology/Bioinformatics/Chemistry/CS/Stats or related).
- 4+ years post-PhD in antibody discovery/engineering/developability data.
- Experience analyzing antibody developability assays (e.g., HIC, AC-SINS, nanoDSF, polyspecificity, viscosity, chemical liabilities).
- ANARCI (or equivalent) + Kabat/Chothia/IMGT numbering knowledge.
- ML validation protocol design for biological sequence data (similarity-aware splits; held-out tests).
Additional Preferences:
- Protein/antibody language model fine-tuning; sequence liability motifs knowledge.
- Federated learning frameworks (NVIDIA FLARE/Flower/OpenFL/PySyft); uncertainty quantification & calibration; PyTorch/ML ecosystem (HF, scikit-learn, RDKit).
- Antibody structure prediction tools (AlphaFold-Multimer/IgFold/ABodyBuilder); regulatory knowledge; MLflow/W&B; publications and technical writing.
Location/Travel: Indianapolis, San Francisco, or Boston; up to 10% travel.