Intended Audience

Biopharma decision-makers, including:

• Discovery and translational research leaders (VPs, directors, program leads) evaluating new targets and disease associations.
• Corporate strategy and portfolio managers assessing pipeline feasibility and competitive landscapes.
• Biotech and pharma executives (CSOs, CMOs) looking for practical experience directing LLMs on scientific data.

No programming, data science, or prior AI background is required.

Problem Statement

Evaluating a potential therapeutic target requires gathering data across multiple domains: human genetics, disease associations, structural druggability, and active clinical trials. Today, these data live in different repositories, databases, and literature sources.

For scientific leaders, running an exploratory target evaluation typically requires submitting requests to data science teams or hiring external consultants, creating delays for preliminary diligence questions.

Large language models can now translate plain-language questions into database queries and analysis steps. When connected to public and internal datasets on Google Cloud (such as BigQuery and DeepMind structural models), models can retrieve genetic evidence, summarize competitive clinical activity, and generate working dashboards in a single session.

Syllabus & AI Competencies Covered

• Directing AI models with scientific context: Structuring prompts, constraints, and domain context so models perform multi-step analysis accurately.
• Querying biological databases with natural language: Using Gemini to query public and proprietary datasets in Google BigQuery without writing SQL.
• Target evaluation workflows: Gathering and synthesizing genetic associations, disease links, and structural druggability predictions.
• Clinical landscape analysis: Searching and summarizing active clinical trials, competitor pipelines, and patent disclosures.
• Fact-checking and citation tracing: Verifying model outputs against source literature and underlying database tables to prevent hallucinations.
• Building interactive summaries: Exporting structured findings into Looker dashboards for team review and diligence reports.

Example Work Product

An interactive target evaluation dashboard in Google Looker.

During the session, each participant selects a therapeutic target and uses Gemini on Google Cloud to query genetic databases, assess druggability, and pull competitive clinical trials. Participants leave with a functioning, source-linked dashboard summarizing the target’s feasibility profile.

Contact