🔬 SiliconMind — Multi-Agent DFT & Semiconductor Validation AI

A domain-specific multi-agent AI system for DFT and semiconductor validation engineers. Ask questions like a senior engineer answers them.

LD7237 Contemporary Computing and Digital Technologies · MSc · Northumbria University

---

What It Does

SiliconMind routes your question to the right specialist agent automatically:

Agent	Handles
⚡ ATPG Agent	Fault models, coverage, test patterns, scan compression
🧠 MBIST Agent	Memory test, March algorithms, repair, BIST architecture
🔍 Scan Debug Agent	Scan chain failures, EDT, silicon bring-up, ATE correlation
🔌 JTAG Agent	Boundary scan, TAP controller, IEEE 1149.1, IJTAG
📰 Research Agent	Live papers from ArXiv and Semantic Scholar (IEEE)
🔬 General Agent	Cross-domain, interview prep, methodology

Built-in tools the agents can call:

• 🔧 search_arxiv — live semiconductor research papers (free, no key needed)
• 🔧 search_semantic_scholar — IEEE + open-access papers
• 🔧 calculate_fault_coverage — computes coverage %, grades it, gives recommendations
• 🔧 generate_debug_checklist — structured checklists for scan, ATPG, MBIST, JTAG, bring-up
• 🔧 fetch_paper_summary — summarise any ArXiv paper from a URL

---

Quick Start (Local)

# 1. Clone
git clone https://github.com/RekhanthReddy/siliconmind.git
cd siliconmind

# 2. Create conda environment
conda create -n siliconmind python=3.12 -y
conda activate siliconmind

# 3. Install
pip install -r requirements.txt

# 4. Run
streamlit run app.py

Get a free API key at console.anthropic.com and paste it in the sidebar.

---

Deploy to Streamlit Cloud (Free, Public URL)

1. Push this repo to GitHub
2. Go to share.streamlit.io → New app → select your repo → app.py
3. Under Advanced settings → Secrets, add:

   ANTHROPIC_API_KEY = "sk-ant-your-key-here"
   

4. Click Deploy — live in ~2 minutes

---

Architecture

User Question
      │
      ▼
 Orchestrator  ─── classifies intent ──► Specialist Agent
      │                                        │
      │                                   Uses Tools:
      │                                   • search_arxiv
      │                                   • search_semantic_scholar
      │                                   • calculate_fault_coverage
      │                                   • generate_debug_checklist
      │                                   • fetch_paper_summary
      │
      ├── ChromaDB (your uploaded PDFs)  ← RAG layer
      └── ArXiv / Semantic Scholar       ← Live research layer

---

Contributing

See CONTRIBUTING.md — add your DFT documents to community_docs/ and open a PR.

---

Tech Stack

Layer	Technology
LLM	Claude Sonnet (Anthropic) with tool use
Vector DB	ChromaDB (local persistent)
Document parsing	pdfplumber, python-docx, python-pptx
Live research	ArXiv API, Semantic Scholar API
UI	Streamlit
Language	Python 3.12