appointat / leagent

A Chatbot of Data Science Expert- Chat with Document(s) using ChatGPT API and Text Embedding

Professional Snapshot

Welcome! I'm a forward-thinking software engineer with a strong focus on user-driven application development and innovation with Large Language Models (LLM). Currently, I'm seeking roles that can benefit from my expertise in domain-specific knowledge bases, prompt engineering, and open-source contributions.

As a core maintainer and contributor at the CAMEL-AI open-source community, I spearhead the design, development, and maintenance of multi-agent systems and interactive agents. In addition, I serve as a backend engineer for Ocademy’s open-source community, harnessing AWS and DynamoDB to drive our systems. I also independently developed a product akin to New Bing that offers embedded links in responses, demonstrating and detailing how to guide AI learning with resourceful links.

My mathematical proficiency is not just academic; it's proven in practice, with recognition as an Outstanding Winner in the Mathematical Contest in Modeling. This underscores my capability in applying theoretical models to real-world challenges.

I'm furthering my skills with a Master's in Computer Science/Engineering in Europe. My background includes enriching internships at ANSYS and Air Liquide and a well-rounded technical skill set in C++, React/JS, Python, and C#/.NET, alongside experience with TensorFlow and PyTorch.

LLM excites me, so what I'm doing or will do (perhaps as a hobby):

1. Enhancing LLM's reasoning abilities, including planning under complex information and solving math problems.
2. Generating and synthesizing data
3. Improving workflow agility for agents (open sourced)
4. Developing a new RAG system with stronger learning capabilities
5. Creating guided agents
6. What is the next

Outside of my professional endeavors, I enjoy keeping up with the latest technological advancements, delving into personal projects, and sharing insights through my technical blogging.

Discover more about my contributions on my GitHub. I'm eager to connect and explore how we can drive innovation together.

Open for Opportunities: I welcome discussions on roles where my unique blend of skills and open-source experience can add value.

leagent's Issues

Chatbot-Directed Traffic to Specific Web Pages

Hello,

I am writing to bring up an issue regarding chatbot-directed traffic to specific web pages. We have observed that the chatbot, in certain circumstances, directs users to specific web pages. While this can be useful in many situations, it poses potential challenges and questions that need to be addressed.

Problem:

There are a few concerns related to this behavior:

Relevance: The chatbot might direct users to pages that are not entirely relevant or may not provide the most up-to-date information.

Security: The chatbot could potentially direct users to insecure or harmful websites. While this is unlikely, given that reputable chatbot services typically have safety checks in place, it's a potential issue that needs consideration.

User Experience: Users might find it disconcerting to be directed away from their current interaction with the chatbot, especially if the transition is not smooth or if the external page doesn't offer a good user experience.

Possible Solutions:

Implement stricter controls on the chatbot's ability to send users to external web pages. This could involve manual approval of suggested links or an improved algorithm for determining link relevance and safety.

Improve the user experience when transitioning to an external page. This could involve opening the link in a new tab or providing a clear warning that the user is about to leave the chatbot interface.

Continually update the chatbot's knowledge base and ensure it has access to the most recent and relevant information, reducing the need to direct users to external websites.

I hope this issue can be considered for future updates. I believe addressing this can enhance user experience and the overall performance of the chatbot.

Thank you for your attention to this matter.

Exposed API Key

https://github.com/Appointat/Chat-with-Document-s-using-ChatGPT-API-and-Text-Embedding/blob/11f2f676ca9f8b780f81bfaf754e3a0a2bebf0fa/Voice%20Chat%20Bot/uiwh.py#L5

Howdy,

Do not expose your API key in your source code.

Instead, consider storing it an environment variable, or using a key management service, or even software such as Vault.
Then be sure to rotate your API key after it is safely stored, or else it may still be recoverable in your commit history.

Also, see Openai's best practices page for api keys for more details:
https://help.openai.com/en/articles/5112595-best-practices-for-api-key-safety

QDRANT_URL and QDRANT_API_KEY

Firstly thank you for making such an interesting chatbot concept. I tried to run your repo but encountered the env problem.
Would you mind giving me the QDRANT_URL and QDRANT_API_KEY?

If it is not safe to share it here, please email me at ([email protected]). Thanks and keep up the good work.