Let’s hear from our Cybersecurity with AI Research intern — Dinki Gupta about their Summer internship 2024 experience.
I’m Dinki Gupta, a final-year student at Sharda University, where I’m pursuing a B.Tech in Computer Science and Engineering with a specialisation in Data Science. Recently, I completed a two-month internship at CRAC Learning, a forward-thinking cybersecurity startup, where I had the opportunity to work as an AI/ML intern. This experience was both challenging and rewarding, offering me the chance to apply my knowledge in a practical setting and gain new insights into the world of AI.
Why CRAC Learning?
CRAC Learning’s focus on cybersecurity initially drew me in, given the growing importance of this field. However, what truly excited me about this internship was the opportunity to work with generative AI (GenAI), an area I had been eager to explore in depth. The combination of cybersecurity and AI presented a unique learning opportunity, making this internship the perfect fit for my interests and aspirations.
My Role and Responsibilities:
During my internship, I was primarily focused on developing a sophisticated conversational chatbot using a Retrieval-Augmented Generation (RAG) pipeline. This project was more than just about building a basic chatbot; it involved creating a system that could answer complex queries based on a custom dataset, manage sessions with memory to maintain context, and even generate and execute Python code on demand. The integration of these advanced features required me to delve deep into generative AI and understand how to effectively combine retrieval mechanisms with AI models to deliver accurate and contextually aware responses.
In the initial phase of my internship, I worked on a phishing mail detection project. This served as a warm-up during my first week, allowing me to familiarize myself with the tools and techniques I would be using throughout the internship. While it was an important introduction, my main focus quickly shifted to the chatbot project, which offered a more complex and engaging challenge.
Technical Details:
During my internship, I developed a sophisticated conversational chatbot using a blend of advanced AI tools and frameworks. Here are some of the key technical aspects of the project:
1. LangChain Framework: The chatbot was primarily built using the `langchain` library, leveraging components like `OpenAIEmbeddings` for embedding creation and `ChatOpenAI` for managing dialogue generation. This allowed for smooth interaction management and the ability to generate contextually accurate responses.
2. Memory and Session Management: One of the critical challenges was implementing effective session management. I used `AgentTokenBufferMemory` to maintain the state of conversations, which allowed the chatbot to remember past interactions and provide more coherent responses across multiple user exchanges.
3. Real-Time Python Code Execution: The chatbot featured the ability to execute Python code in real-time, a capability made possible through the integration of `PythonREPLTool`. This tool enabled users to input Python commands directly into the chatbot, which then executed the code and returned the output instantly.
4. Custom Prompt Design: I carefully crafted prompts to ensure the chatbot could interpret user inputs accurately. The prompts were designed to handle a wide range of queries, from simple questions to complex requests involving code execution.
5. Session Persistence and File Management: To ensure that conversations could be resumed or reviewed, session data was stored and retrieved from files. This feature was crucial for maintaining the continuity of interactions and providing users with a seamless experience.
Conclusion:
My internship at CRAC Learning has been a pivotal experience in my academic and professional journey. The opportunity to work on a sophisticated chatbot, combined with the initial exposure to phishing mail detection, has provided me with invaluable skills and knowledge that I am eager to apply in future endeavors.
I am deeply grateful to the CRAC Learning team for their guidance and support throughout this internship. As I move forward in my studies and career, I am excited to continue exploring the intersection of AI and cybersecurity, applying the lessons I’ve learned to new challenges.
This internship has been more than just a learning experience; it has been a significant stepping stone toward my future goals in the field of AI and data science. I look forward to what lies ahead!
Comentarios