FundaGPT: how I leveraged LLMs to fully automate the search for a house using langchain, GCP and OpenAI
Finding a home is difficult. Especially in bigger cities such as Eindhoven and Amsterdam, it is becoming more- and more tricky to buy a house. Therefore, we came with a solution: FundaGPT. FundaGPT is your personal broker. It notifies you when relevant houses come online, it can answer any additional questions that you have, and it can even send e-mails to schedule a visit!
During this talk, we will dive deeper how Langchain’s GenAI Agents are used to achieve this. We will take a brief look at theoretical concepts, and how they are implemented in the Langchain framework. What are these agents, and what are common challenges during development? Moreover, how can you cost-effectively deploy them in GCP? By using a combination of Cloud PubSub, Cloud Functions, Datastore, Terraform and Langchain, we will show how this personal broker can be deployed in no-time, at minimal cost. This talk aims to get you started with writing your own agents, such that you can use them to automate all kinds of tasks!
Xebia | Data
Sander van Donkelaar
Machine Learning Engineer