JakeAI | Devpost

Homepage
Title

Inspiration

We were inspired to create JakeAi by the desire to help people live better, healthier, and more organized lives. We recognize that people are busy and often don't have the time or energy to figure out how to do everything on their to-do list. JakeAi can help take some of the burden off of people by providing guidance and support on a wide range of topics. We were also inspired by the potential of artificial intelligence to make people's lives better. We believe that JakeAi has the potential to be a valuable tool for people of all ages and backgrounds. He can help people learn new things, save time and money, and make their lives easier and more manageable.

What it does

JakeAI is your all-in-one solution for hassle-free living. This AI-powered chatbot is designed to provide expert guidance and support in three key areas: vehicle maintenance, home upkeep, and life in general. Whether you're looking for advice on keeping your car in top condition, maintaining a comfortable and safe home environment, or seeking general life tips, JakeAI has you covered.Our app doesn't just solve your problems, it also helps you decide how State Farm would help you if you had purchased a particular policy.

How we built it

We have developed an application named JakeAi, featuring a frontend built on Streamlit. The core functionalities include the integration of LangChain, LLM (ChatGPT-3.5 Turbo), and Vector Databases. The JakeAi chatbot has been meticulously trained on an extensive dataset comprising case studies, policies, and various services offered by StateFarm

Challenges we ran into

Data Quality and Precision: Securing high-quality data for accurate results proved challenging. The iterative refinement of our dataset highlighted difficulties in ensuring completeness and relevance. Integration Complexity: Harmonizing LangChain, LLM (ChatGPT-3.5 Turbo), and Vector Databases introduced integration complexities. Fine-tuning and troubleshooting were pivotal for seamless collaboration among these technologies. Training Model Dynamics: Training the chatbot for StateFarm's services presented multifaceted challenges plus also training our chatgpt model to help customers was a problem since they both were creating different sessions. User Experience Optimization: Despite choosing Streamlit for the frontend, optimizing the user experience posed challenges. Striking the right balance between functionality and simplicity demanded constant refinement. Security and Compliance: Ensuring data security and compliance with industry regulations, particularly in the insurance sector, demanded rigorous attention throughout the development process.

Accomplishments that we're proud of

Our chatbot provides good results that anyone who is using the assistant will be able to benefit from it and save time rather and going through the internet.

What we learned

We learned a lot about Langchain technologies, vector database, stramlit, LLM(chatgpt-3.5 turbo) and aloot of debugging. We also did learn about on how to use react but unfortunately we ran into some problems, thats why we switched it to streamlit.