• 

  SafeLabs making labs safer

• 

  The Homepage of our application

• 

  The SafeLabs Document Retrieval and Conversational Agent. Gives checklists, summarizes, gives guidance and more.

• 

  Voice-based Guide, The SafeLabs Scribe Assistant that retrieves data and provides important compliance and safety guidance.

• 

  The entry point for the instant knowledge check and quiz generation.

• 

  The instant knowledge check/quiz questions

• 

  The result of the quiz

• 

  About Page

SafeLabs: Ensuring Safety at Research Labs

SafeLabs strives to ensure safety and compliance at Research and Science Labs. It's the first stepping stone on a comprehensive AI platform for research institutes.

Problem Statement

Research Labs working on cutting edge sciences have many compliances and safety procedures to follow. It is often tedious to achieve this, but a single mistake can take a life, like Sheri Sangji in 2008.

big>How Might We ensure that compliance and safety procedures are followed? We take the help of AI and Large Language Models (LLMs)!

Why Us

Who better than the scientists themselves to identify and solve problems. Our team consists of three Berkeley Lab scientists who have extensive experience in the field and are intricately aware of the problems faced by the community.

Our Solution

We have developed a comprehensive AI platform that uses Large Language Models (LLMs) to answer queries with citations. The platform is designed by scientists for scientists. We make extensive use of LLM APIs, technologies like Streamlit, Text-to-Speech, Speech-to-Text to deliver a pleasant and seamless experience.

Tech Stack: Python, Flask, Streamlit, OpenAI API, Langchain API, Groq Speech-to-Text API, and more.

Challenges

Ensuring the results are truthful, accurate and the platform is fast.

Accomplishments

We are proud of delivering an easy to use and understand platform that answers complex queries in natural text/voice. We are proud of the extensibility and future potential of the platform.

What We Learned

Effective usage of APIs, prompt-tuning, integration of disparate technologies and in-depth understanding of the problems involved.

What's Next - Zooming Out

We envision SafeLabs as a first step towards a comprehensive AI application suite for research labs and institutes. The potential is immense and the disruption it can bring about is significant.

Potential Customers

Academic and teaching labs, Chemical Companies, OSHA, Environment, Health and Safety departments, Waste disposal companies, US government agencies like Environment Protection Agency

Built With

• flask
• langchain
• largelanguagemodels
• openai
• python
• streamlit