Skip to content

Latest commit

 

History

19 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 

Repository files navigation

PharmaIntel Enterprise Dossier Cockpit

PharmaIntel 🧬

Agentic AI Platform for Drug Repurposing & Pharma Intelligence

PharmaIntel is a full-stack, agentic AI–powered research platform designed to accelerate drug repurposing, innovation research, and strategic decision-making in the pharmaceutical domain.
It orchestrates multiple AI agents to analyze clinical trials, patents, market data, trade intelligence, and internal documents, delivering decision-ready insights through a modern web dashboard.


🚀 Key Features

  • 🔍 Natural Language Research Queries
  • 🤖 Agentic AI Orchestration (LangGraph-based)
  • 🧪 Clinical Trials Intelligence (Mocked ClinicalTrials.gov)
  • 📜 Patent Landscape & Freedom-to-Operate Analysis
  • 📊 Market & EXIM Trade Insights
  • 📁 Internal Document Summarization
  • 📈 Visual Dashboards & PDF Report Generation
  • 🧠 Explainable, Auditable AI Outputs

🧠 System Architecture Overview

High-level Flow:

PharmaIntel Architecture

Architecture Overview

PharmaIntel follows a layered, agent-driven architecture designed for scalability, explainability, and modular expansion.
The system orchestrates multiple domain-specific AI agents through a centralized master agent to deliver decision-ready pharmaceutical intelligence.


🔹 User Interface Layer

  • Web-based React dashboard for natural language research queries
  • Real-time visualization of insights, charts, and reports
  • Agent execution status, logs, and traceability
  • Modular UI for future feature expansion

🔹 Orchestration Layer (Master Agent)

  • Interprets and parses user intent
  • Decomposes complex queries into domain-specific tasks
  • Manages workflow state using LangGraph
  • Coordinates parallel execution of worker agents
  • Aggregates intermediate results for synthesis

🔹 Intelligence Layer (Worker Agents)

Each agent operates independently and in parallel to maximize efficiency:

  • Clinical Trials Agent – Analyzes trial phases, outcomes, and trends
  • Patent Landscape Agent – Evaluates patent status, lifecycle, and FTO risks
  • Market Intelligence Agent – Assesses market size, growth, and competition
  • EXIM Trade Agent – Extracts import–export and trade flow insights
  • Web Intelligence Agent – Mines scientific publications and guidelines
  • Internal Knowledge Agent – Summarizes internal documents and reports

🔹 Insight Synthesis & Reporting Layer

  • Cross-domain insight correlation and validation
  • Confidence-weighted result aggregation
  • Automated generation of:
    • Interactive dashboards
    • Visual analytics (charts, heatmaps)
    • Decision-ready PDF research reports

🔹 Infrastructure & Deployment

  • Containerized services using Docker
  • Modular, cloud-ready deployment
  • Designed for scalability across AWS, Azure, or GCP
  • Secure and auditable execution flow

Design Philosophy:
Parallel intelligence, deterministic orchestration, and explainable outputs — built for real-world pharma decision-making.


🏗️ Tech Stack

Frontend

  • React.js
  • TypeScript
  • Modern UI components & data visualizations
  • Agent execution status & logs

Backend

  • Python
  • FastAPI
  • LangGraph for stateful agent orchestration
  • Modular AI agent architecture

AI & Data

  • LLM-based reasoning
  • Mocked APIs simulating:
    • ClinicalTrials.gov
    • Patent databases (USPTO-like)
    • Market intelligence platforms
    • EXIM trade data
  • Internal document processing

Infrastructure

  • Docker-based modular setup
  • Cloud-ready architecture

📂 Project Structure

pharmintel/
│
├── frontend/ # React + TypeScript frontend
│ ├── components/
│ ├── pages/
│ ├── services/
│ └── visualizations/
│
├── backend/ # FastAPI backend
│ ├── agents/ # Individual AI agents
│ ├── orchestration/ # LangGraph workflows
│ ├── api/ # REST endpoints
│ ├── services/
│ └── utils/
│
├── docker/ # Docker configuration
├── README.md
└── requirements.txt

📊 Visualizations & Outputs

The platform delivers insight-rich, decision-ready outputs through interactive and automated visualizations, including:

  • Bar Charts, Donut Charts, and Heatmaps for:
    • Patent lifecycle distribution
    • Market opportunity analysis
    • Clinical trial phase breakdown
  • Agent-wise execution timelines for transparency and traceability
  • Auto-generated PDF research reports summarizing cross-domain insights

🔐 Assumptions & Constraints

Assumptions

  • Mock data sources are acceptable for MVP and hackathon implementations
  • User queries are molecule-centric or indication-centric
  • Human validation is available for high-impact or critical decisions

Constraints

  • Paid real-world data sources are not directly integrated
  • LLM context length and reasoning limitations
  • Outputs must remain explainable, auditable, and transparent

🎯 Use Cases

  • Drug repurposing opportunity analysis
  • Patent landscape and Freedom-to-Operate (FTO) assessment
  • Market entry strategy and trade intelligence
  • End-to-end pharmaceutical research acceleration
  • Strategic decision support for R&D and innovation teams

🧪 Current Status

  • ✅ MVP ready
  • 🔄 Mocked data integrations in place
  • 🚧 Real-world data connectors planned

🌱 Future Enhancements

  • Live API integrations (ClinicalTrials.gov, patent databases)
  • Role-based access control (RBAC)
  • Advanced financial and commercial modeling
  • Multi-language query and report support
  • Cloud-native deployment (AWS / Azure / GCP)

This is a Next.js project bootstrapped with create-next-app.

A Multi-Agent System (MAS) designed to automate the Hypothesize-Test-Pivot cycle in pharmaceutical R&D.

Architecture

This solution implements a cyclic LangGraph workflow:

  1. Planner: Formulates a scientific hypothesis (e.g. "Repurpose Drug X for Y").
  2. Test (Agents): Parallel execution of specialized agents:
    • ClinicalAgent: Analyzes trial data and safety signals (ClinicalTrials.gov).
    • PatentAgent: Checks for freedom-to-operate and white space (USPTO).
    • IQVIAAgent: Analyzes market size and CAGR trends.
  3. Verifier: Evaluating if the findings support the hypothesis or if there are blockers (e.g., toxicity).
  4. Failure Analysis & Pivot: If verified as "Failure", the AI analyzes why and suggests a Pivot (e.g., "Switch to Inhaled formulation"), restarting the loop.

Directory Structure

  • backend/: FastAPI + LangGraph + Google Gemini
  • frontend/: Next.js 14 + TailwindCSS (Dark Mode Console)

Getting Started

Prerequisites

  • Google API Key: Put in backend/.env.

Quick Start

./start.sh

Manual Start

Backend:

cd backend
uvicorn app.main:app --reload

Frontend:

cd frontend
npm run dev

Open http://localhost:3000.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages