Position Summary

We are building LLMs for biological research and are seeking a Research Engineer / Scientist to lead post-training experiments that improve reasoning, tool use, and reliability.

The primary focus will be on LLMs and post-training—spanning training data and reward design through optimization and evaluation. You will collaborate with ML engineers and domain scientists to translate promising methods into measurable model improvements. Prior biology experience is welcome but not required.

Key Responsibilities

• Develop and iterate on SFT, preference optimization, and reinforcement learning recipes.

• Build scalable pipelines for rollout generation, reward computation, and policy training.

• Improve multi-step reasoning, tool use, and performance on complex scientific tasks.

• Develop synthetic data, distillation, and data filtering methods.

• Design rewards and evaluations; diagnose reward hacking, training instability, and capability regressions.

• Run rigorous experiments and bring successful methods into production.

Professional Experience and Qualifications

• Required:

• Strong foundations in deep learning, Transformers, and language modeling.

• Hands-on experience training or post-training LLMs, with depth in SFT, DPO, PPO, GRPO, or related methods.

• Proficiency in Python and PyTorch; comfortable debugging distributed training.

• Ability to independently design experiments, analyze model behavior, and demonstrate reliable improvements.

• Preferred:

• Experience with large-scale RL, reasoning models, or tool-using agents.

• Familiarity with frameworks such as veRL, OpenRLHF, vLLM, or SGLang.

• Experience with reward modeling, synthetic data, or model distillation.

• Interest in biology, biomedical research, or scientific discovery.

Position Summary

We are seeking a scientist-leader to build and lead our Antibody Engineering platform as a horizontal, cross-functional engine serving all therapeutic programs. Reporting to the CSO, you own the end-to-end antibody engineering workflow—across monoclonal, bispecific and multispecific formats—raising the pipeline-wide success rate rather than driving any single program to PCC. You are an antibody therapeutics expert with a proven track record in therapeutic antibody development, particularly in resolving developability challenges through antibody engineering and in designing and developing bispecific and multispecific antibodies. You will codify the playbook, run a tight AI-to-wet-lab loop, and build a high-performing team in an AI-integrated discovery company.

Key Responsibilities

• Platform Strategy

Set the multi-year roadmap; own the technology stack, stage-gates, and quality standards across library/display-based discovery, affinity maturation, Fc, bispecific and multispecific engineering, mammalian expression, and developability. Make developability a design principle rather than a late-stage check: own the assessment and the engineering solutions that fix liabilities (expression, aggregation, solubility, viscosity, stability, PK, immunogenicity) from the antibody engineering side. Drive bispecific/multispecific format choice—how to assemble each format and which format to select for the target biology and manufacturability.

• AI × Wet-Lab Loop

Partner with the in-house AI team to embed predictive design and in silico screening into experimental workflows, and curate internal data assets so the loop keeps improving.

• Cross-Project Enablement

Serve as central technical authority: chair key reviews (antigen, hit, lead, PCC), triage engineering problems, and allocate platform resources across programs.

• Team & Ecosystem

Build and lead the wet-lab team (library/display, HTS, molecular biology, expression, analytics); set CRO/partner strategy and pursue academic collaborations that extend internal capabilities.

• IP & Knowledge

Identify patentable platform innovations; drive knowledge management and external scientific engagement.

Professional Experience and Qualifications

Required:

• Ph.D. in Antibody Engineering, Immunology, Molecular Biology, or related field, with 5+ years of therapeutic antibody R&D experience.

• Hands-on depth across discovery and engineering: library/display-based screening, lead isolation, affinity maturation, Fc, bispecific and multispecific engineering, mammalian expression, developability.

• Strong developability engineering: a proven record of solving developability problems (aggregation, solubility, viscosity, stability, expression, PK, immunogenicity) from an antibody engineering angle.

• Direct, hands-on bispecific/multispecific antibody experience: a proven track record of designing, assembling and advancing bispecific/multispecific molecules across Fc-based and fragment-based formats, with the judgment to select the right format for each target’s biology and manufacturability.

• Track record of technical leadership on discovery programs, including go/no-go input at key gates.

• Open, collaborative mindset toward AI-driven R&D.

• Strong cross-functional influence; fluent in English and Chinese.

Preferred：

• Hands-on experience building or scaling display platforms (phage, mammalian cell, ribosome) and HTS workflows.

• Prior experience standing up a platform/function from scratch in early-stage biotech.

• A record of leading multiple bispecific/multispecific programs through advanced discovery gates (lead to PCC), including complex formats and manufacturing-friendly design.

Why Join Us

• Own the horizontal engine that powers the whole pipeline; direct line to the CSO.

• Work at the sharp edge of AI-integrated antibody discovery with genuine compute–experiment loops.

• Build a team in a science-first, low-politics environment with real autonomy.