<Dev Sensei>

Dev Sensei is a productivity platform tailored for developers, empowering them to design a system with a suitable & pertinent tech stack for developing applications, websites, or software services.

Comment 1

Inspiration

Our aspiration was to create an AI-driven platform designed to mentor and support novice developers at the beginning of their learning path. We wanted to leverage advanced language models to provide personalized and contextual recommendations to developers.

What it does

A productivity platform tailored for budding developers, empowering them to plan and construct a system with a suitable and pertinent technology stack for the development of applications, websites, or software services.

  1. A comprehensive answer to the user's query would include an explanation of the technology choices made for the tech stack response.
  2. Provides a system design illustration for visual clarity and enhanced comprehension.
  3. Provides curated resources to help them learn.

How we built it

Backend:- Node.js, Flask API Frontend:- React JS AI:- GPT 3.5 Turbo, ChromaDB (experimented) DB:- MongoDB Atlas Document Store Hosting:- Google Cloud Compute (GCP) CI/CD: Github Actions Proxy Manager:- Nginx

Challenges we ran into

  1. Integrating different AI models and databases proved difficult initially. We experimented with various architectures before settling on the final tech stack.
  2. Scaling the platform to handle increased information from AI Models without compromising performance was also a key challenge.
  3. GPT 3.5 API stability issues.

Accomplishments that we're proud of

We successfully built an end-to-end platform that can understand developers' needs and provide tailored solutions. The platform can intelligently recommend relevant technologies, resources, and next steps based on developers' queries.

What we learned

In our evaluation of AI backends, we conducted extensive testing on various Large Language Models (LLMs) such as GPT, PaLM, Claude, Phind.AI, Llama, and Dolly to assess their performance and the quality of their responses. It became evident that both GPT and Claude consistently delivered superior response quality compared to the other models. As a result, we opted for GPT 3.5 Turbo due to its faster response times, surpassing Anthropic's Claude AI.

What's next for

Provide code snippets and templates to help developers quickly get started with implementation. This could include starter projects, boilerplate code, and code examples for common tasks. Offer mentorship & coaching options by connecting junior developers with more experienced mentors.

Built With

Share this project:

Updates