× Strengths About Experience Projects Skills Contact
Senior Researcher · Hansoh Biomedical · Shanghai

Jianzhong Chen Bridging Biologics R&D with AI Intelligence

6+ years of end-to-end CMC experience across mAb, ADC, and fusion proteins. Developer of PDassistant — an LLM-powered CMC intelligence agent that automates regulatory workflows, formulation analysis, and project monitoring.

Explore PDassistant View Full Experience
6+
Years CMC R&D
10+
Drug Programs
75%+
ML Accuracy
130k
Token Window
Core Strengths

Four Pillars of Expertise

Where deep domain science meets cutting-edge AI engineering.

AI Agent Design

Architect of PDassistant — a production-grade CMC intelligence agent that orchestrates RPA (TagUI), Office automation (win32com), and large language models (Qwen / Gemini) into end-to-end research workflows. The system handles project monitoring, FDA label parsing, literature mining, and daily regulatory intelligence briefings with a 130k-token context window.

Multi-Agent LLM RPA · TagUI win32com Qwen / Gemini
CMC Formulation Expert

End-to-end formulation development across mAb, ADC, fusion proteins, and asymmetric bispecifics. Expert in high-concentration SC formulation (20 → 120 mg/mL), lyophilized dosage forms, co-formulation with hyaluronidase, and full IND/BLA CMC dossier authorship.

mAb / ADC IND / BLA Filing HC ≥120 mg/mL
Predictive ML Modeling

Built the internal Drug Developability Platform using Random Forest models on cleaned experimental datasets. Delivers >75% accuracy for viscosity and aggregation prediction, cutting 1-month development cycles.

Viscosity Model78%
Aggregation Model75%
Prompt Engineering

Specializes in domain-adapted prompt design for CMC contexts — translating complex regulatory logic, FDA/EMA label structures, and scientific literature into precise LLM instructions. Built the Smart-Crop PDF extraction algorithm and multi-step reasoning chains for project risk analysis. Transforms manual expert workflows into automated, auditable AI pipelines.

Chain-of-Thought Smart-Crop™ RAG Design Regulatory Parsing
Background

About Jianzhong

I'm a Senior Researcher (Scientist III) at Hansoh Biomedical in Shanghai, specializing in biologics CMC formulation and process development. With 6+ years spanning mAb, ADC, and complex fusion protein programs, I bring both deep bench science and a systems-level perspective to drug development.

What sets me apart is bridging the gap between traditional lab expertise and AI-driven automation. I built PDassistant from scratch — a multi-agent LLM system that replaced hundreds of hours of manual regulatory research and project reporting with automated, real-time intelligence workflows.

My academic roots are in pathogenic microbiology at CAS, where I designed covalent anti-influenza compounds via structure-based CADD and developed liposomal drug delivery systems — giving me a uniquely interdisciplinary foundation across chemistry, biology, and computational science.

Hansoh Biomedical · Scientist III Led HC formulation platform (>120 mg/mL), ADC lyophilization, and built PDassistant in production.
Henlius Biotech · Researcher II Reduced protein aggregation by 30%, built accelerated stability ML models, managed DOE campaigns.
M.S. · CAS Institute of Microbiology Designed covalent anti-influenza agents via CADD; liposomal formulation to enhance oral bioavailability.
Lab Infrastructure & Knowledge Systems Built excipient databases, structured budget templates adopted department-wide, and managed million-RMB equipment procurement.
Career Timeline

Professional Experience

A decade of progressive expertise across biologics research, process development, and AI-driven automation.

Sep 2023 — Present
Senior Researcher · Scientist III
Hansoh Biomedical Co., Ltd · Process Development Department · Shanghai
Current Role
High-Concentration Subcutaneous Formulation Platform: Led end-to-end development of SC mAb formulations from 20 mg/mL up to >120 mg/mL co-formulated with hyaluronidase. Executed full process characterization including risk assessment, formulation robustness studies, and scale-down model establishment for regulatory submissions.
ADC Formulation & Lyophilization Development: Developed liquid and lyophilized formulations for multiple ADC molecules spanning different payloads (MMAE, DM1, and novel linkers) and DAR values (2–8). Optimized cryoprotectant systems, collapse temperature margins, and reconstitution profiles for clinical-stage programs.
PDassistant — CMC AI Agent (Developer): Single-handedly designed and deployed a multi-agent LLM system integrating RPA (TagUI), Python, and win32com to automate: (i) Stage 1–5 project Gantt monitoring and delay risk analysis; (ii) Smart-Crop algorithm for FDA/EMA label extraction and IND precedent mining; (iii) automated PubMed/Google Scholar literature intelligence with structured Excel export; (iv) daily Outlook email summarization.
Drug Developability Platform: Designed and maintained a machine learning prediction pipeline (Random Forest) for high-concentration viscosity and aggregation propensity. Data cleaning, feature engineering, and model validation on internal experimental datasets; achieved >75% predictive accuracy, saving approximately one month per development cycle.
Excipient Intelligence Database: Compiled and curated a proprietary excipient database cross-referencing regulatory precedents, compatibility data, and novel excipient candidates. Reduced formulation screening time and contributed to multiple patent filings.
Laboratory Infrastructure Lead: Directed layout planning for a new analytical laboratory; managed vendor selection, electrical/plumbing coordination, SOP authoring, and procurement of million-RMB equipment (e.g., DLS, SEC-HPLC, viscometers). Established department-wide structured budget templates adopted across the PD function.
Formulation Development Portfolio @ Hansoh
HC-SC + Hyaluronidase
Sep 2023 – Apr 2024 · SC; 20 mg/mL → 120 mg/mL co-formulated with rHuPH20. Full process characterization & CDE submission package.
ADC Liquid & Lyophilized
Feb 2024 – Jun 2025 · Multi-payload ADC formulations (MMAE, novel warheads); DAR 2–8 stability comparability; cryoprotectant optimization.
HC SC mAb Characterization
May 2025 – Present · Process characterization for HC SC mAb; robustness study design, scale-down modeling, FMEA-guided risk matrix.
IND/BLA CMC ADC mAb HC-SC PDassistant Python · RPA Random Forest
Jul 2020 — Sep 2023
Researcher II
Henlius Biotech, Inc. · Product Development Department · Shanghai
3 Years
mAb Biosimilar & Innovative Drug Formulation: Led formulation development for multiple liquid and lyophilized mAb programs (biosimilars and NCEs). Executed accelerated and long-term stability studies, designed DOE matrices for buffer/excipient screening, and authored CMC sections for CDE IND submissions.
Protein Aggregation Reduction — 30% Achievement: Applied systematic DOE and statistical analysis (JMP) to identify critical formulation variables driving sub-visible particle formation and high-molecular-weight species. Reformulation strategy resulted in a >30% reduction in aggregation across stressed conditions.
Asymmetric Bispecific Antibody Formulation: Pioneered formulation strategy for a complex asymmetric bispecific antibody (knob-into-hole architecture). Applied MOE molecular modeling to assess developability and guide targeted excipient selection for charge heterogeneity and conformational instability.
Accelerated Stability Prediction Model: Built an early-stage machine learning model to correlate accelerated stability readouts (40°C/4-week SEC, DLS) with real-time stability outcomes, enabling faster go/no-go decisions in formulation screening.
Tech Transfer & CDMO Coordination: Managed formulation tech transfer to external CDMO partners. Authored transfer documentation, defined critical quality attributes, and tracked process equivalence across scale-up batches.
Biosimilar CMC Bispecific Ab DOE · JMP Tech Transfer MOE Modeling
Sep 2017 — Jun 2020
M.S. Researcher in Pathogenic Microbiology
CAS — Institute of Microbiology · Key Lab of Pathogenic Microbiology · Beijing
Graduate
Covalent Anti-Influenza Drug Design: Applied structure-based CADD (MOE, Schrödinger) to design irreversible covalent inhibitors targeting influenza neuraminidase. Synthesized a library of Michael acceptor and epoxide-based warhead analogues and characterized their antiviral potency and mechanism of action via enzymatic and cell-based assays.
Liposomal Drug Delivery Research: Developed liposomal formulations of antiviral compounds to modulate log P and improve oral bioavailability. Characterized particle size, zeta potential, encapsulation efficiency, and in vitro release profiles; evaluated pharmacokinetic improvements in rodent models.
CADD · MOE Organic Synthesis Liposomal Drug Delivery Virology
Sep 2012 — Jun 2016
B.Eng. in Bioengineering
Beijing Union University · Faculty of Robotics & Automation
Undergraduate
Featured Work

PDassistant & Platform Projects

Production systems that replaced manual workflows with intelligent automation.

PDassistant
v1.0 · CMC Intelligence Agent

A production-grade multi-agent system purpose-built for CMC research workflows. PDassistant integrates RPA automation (TagUI), Microsoft Office backend (win32com), and frontier LLMs (Qwen 2.5-72B / Gemini 1.5 Pro) into a unified orchestration layer. It replaces fragmented manual processes with auditable, repeatable AI pipelines — handling everything from Gantt chart delay analysis to FDA label parsing and competitive intelligence briefings.

The system's Smart-Crop algorithm addresses the fundamental challenge of extracting structured data from long-form PDF regulatory documents — automatically identifying formulation sections, excipient tables, and dosing information from FDA/EMA labels, then outputting structured summaries to Excel for downstream analysis.

Python TagUI RPA win32com Qwen 2.5-72B Gemini 1.5 Pro Smart-Crop™ Production
System Capabilities
Context Window
130,000 Tokens
Project Monitoring
Stage 1–5 Real-time
PDF Parsing Engine
Smart-Crop™ Algorithm
Literature Mining
PubMed + Google XPath
Email Intelligence
Outlook Direct Connect
Model Routing
Dynamic LLM Switching
Deployment Status
Operational ✓
Drug Developability Platform

Random Forest ML pipeline for viscosity and aggregation prediction from sequence/formulation features. >75% accuracy. Internal platform, Hansoh 2023–present.

Excipient Intelligence Database

Proprietary regulatory intelligence database cross-referencing excipient precedents, compatibility profiles, and novel excipient candidates. Resulted in multiple patent applications.

Covalent Anti-Influenza Program

CADD-guided design and organic synthesis of covalent neuraminidase inhibitors. Liposomal reformulation improving log P and oral PK. CAS Institute of Microbiology, 2017–2020.

Expertise Map

Skills & Tools

CMC Domain Expertise
mAb Formulation ADC Development Fusion Protein Bispecific Ab HC-SC Formulation Lyophilization Hyaluronidase Co-form. Accelerated Stability DOE / JMP HPLC / SEC DLS / Viscometry FMEA Tech Transfer IND / BLA Filing CDE / FDA CMC Process Characterization CDMO Management Scale-down Modeling MOE Molecular Modeling Organic Synthesis Liposomal Delivery
AI / Tech Stack
Python LLM API (OpenAI/Qwen) Gemini API RPA · TagUI win32com Random Forest XGBoost Neural Networks Scikit-learn Pandas / NumPy XPath Scraping PDF Parsing Prompt Engineering RAG Architecture Agent Orchestration JMP Excel Automation Regex Engine Git CADD · MOE
Get In Touch

Let's Build Something
at the Intersection of Biology & AI

Open to collaborations in AI-enabled biologics research, CMC automation consulting, and intelligent formulation platforms. Always interested in what's next.

cjzmyself@live.com +86 188-1150-3325
Shanghai, China