Preclinical R&D Expert – ADC & Bispecifics

What you would do

• Review and annotate preclinical R&D data, model outputs, and scientific documents from ADC and bispecific programs
• Label and classify experimental data so models learn which findings drive development decisions
• Evaluate AI-generated recommendations on compound design, modality choice, and target selection for scientific soundness
• Assess model reasoning about lead optimization, risk mitigation, and translational biology decisions
• Define quality standards and structural requirements for pharmaceutical datasets that train AI systems

Who they want

• Hands-on preclinical R&D experience at a company developing its own drug assets, with programs advanced to the clinic
• Direct experience designing, engineering, or optimizing ADC and/or bispecific (or multispecific) antibody therapeutics
• Experience spanning discovery biology, medicinal chemistry, antibody/protein engineering, computational methods, translational pharmacology, or CMC/manufacturing
• 5+ years industry R&D experience (flexible depending on depth in ADC or bispecific expertise), ideally with lead-optimization or IND-enabling work
• Part-time engagement at approximately 10 hours per week; experience at mid-sized or emerging biotech preferred over large pharmaceutical company backgrounds

Main skills

What the interview asks about

1. 1.ADC design and optimization trade-offs

   ADC development involves balancing payload potency, linker stability, antibody affinity, and immunogenicity. You need to evaluate whether AI reasoning captures these interdependent constraints.

   For example: “Design an ADC for a solid tumor. Explain payload, linker, and antibody choices, weighing potency against off-target toxicity and manufacturability. What must an AI model understand to reason through these trade-offs?”

2. 2.Bispecific antibody target pair selection

   Choosing the right target pair for a bispecific involves assessing target biology, co-expression patterns, and whether the pairing strategy actually creates therapeutic advantage. AI must reason about mechanism design.

   For example: “AI recommends a bispecific targeting two checkpoint proteins on T cells for enhanced activation and killing. What biological, manufacturing, and clinical factors should it have considered but may have missed?”

3. 3.Lead candidate selection and progression

   Programs make go/no-go decisions between candidate molecules based on a constellation of data: potency, selectivity, PK, off-target risks, synthetic tractability. AI must reason about this complexity.

   For example: “Compound A has superior engagement but poor stability; B has the opposite profile. Describe your lead-selection logic, additional experiments needed, and how an AI model should reason through risk and feasibility.”

4. 4.Translational pharmacology reasoning

   Bridging preclinical to clinical requires understanding PK-PD relationships, dose-response assumptions, and how animal models do and do not predict human safety and efficacy.

   For example: “Design an annotated dataset for translational PK-PD reasoning with examples of successful and failed rodent-to-human translation. What annotations teach when to trust preclinical predictions?”

5. 5.Manufacturing and feasibility assessment

   Many brilliant compounds fail development because they're impossible to manufacture at scale, too expensive, or require novel processes. Real scientists factor this into design.

   For example: “AI recommends a novel ADC linker with excellent stability but prohibitive manufacturing cost at scale. How would you annotate this to teach the model about feasibility constraints in therapeutic design?”

6. 6.Interpreting and using preclinical data gaps

   No study is perfect. Programs proceed with incomplete data by accepting uncertainty and mitigating risk. AI needs to learn which data gaps matter and which are acceptable.

   For example: “A bispecific has strong in vitro engagement but limited in vivo data in one xenograft. What would you need before IND studies? What's missing from the model's data interpretation and risk assessment?”

A task you may get

Review three hypothetical ADC or bispecific program decisions. For each, annotate what data matters, what the model should consider, and what reasoning gaps indicate insufficient drug-discovery understanding.

How to prepare

• Prepare a case study from your R&D experience: a lead candidate, optimization work, decision to progress or deprioritize, and influencing factors
• Compile a list of preclinical-to-clinical failures or surprises you've witnessed, and what biomarkers or experiments might have predicted them
• Review current ADC or bispecific programs from published literature and be ready to discuss what makes them scientifically sound or risky based on their disclosed data
• Gather examples of how target biology, manufacturing constraints, and cost considerations shaped your program strategy in ways non-scientists might underestimate

The facts

Pay

$60–100/hr

Hours

Hourly, 40 hours a week

Where

Remote

Field

Life, Physical, and Social Science

Posted

8/12/2026

Places left

4

We wrote this page from the public Mercor listing. It may be out of date, so read the full posting before you apply.