Update Contributions (Axesh Patel)

documentation/T1_2024/Group_Tasks_and_Handover_Documents/DataBytesReport.md

@@ -405,6 +405,7 @@ height="4.840277777777778in"}
 | Nicholas Lane   | Write contributions here ! <li>I developed a NLP model that uses transfer learning to classify transactions using a BERT transformer.</li><li>I wrote code to clean and preprocess text data, encode class labels, then tokenized transaction descriptions using BERT tokenizer, convert to BERT input format.</<li><li>Created a function to create datasets, create datasets for training, validation, and testing.</li><li>Loaded in the Pre-trained BERT Model: Load pre-trained BERT model for NLP classification, then added an output layer for class prediction.</li><li>Compiled and train the BERT model with training dataset and fine tuned the model and finally evaluated the performance of the model.</li><li>Updated the code to include the DolFin colour code format, and resubmitted the code</li><li>I development of a Deep Neural Network model to identify and classify fraudulent bank transactions.</li><li>I searched for a suitable dataset, that would contain the information that DolFin would be able to access through the Open Banking platform.</li><li>I write code to clean and preprocess the various data types and prepare them for the deep learning model.</li><li>I split the data into training and testing datasets which were then made into TensorFlow datasets, which had been shuffled and prefetched.</li><li>I developed a model and added regulation to improve the model’s generalizability and reduce overfitting to the training data.</li><li>>The model was compiled and then evaluated, I also updated the colours to the Dolfin colour format and then submitted the code as a .py file.</li><liAttended product owner meetings and team meetings</li><li>Assist with handover documentation and presentation slides.</li>|
 | Junkai Jiang | Write notes about how to set up Dolfin_new<br/>Set up Dolfin_new GitHub repository<br/>Develop JWT service for user authentication(Dolfin_new)<br/>Develop Basiq API service(Dolfin_new)<br/>Develop database service(Dolfin_new)<br/>Review the pull request by Deepak: Optimization of the clear transaction function.<br/>Review the pull request by Sagar: Email verification function.<br/>Review the pull request for rebuilding the login route and update the login page<br/>Review the pull request for rebuilding the dashboard<br/>Set up Dolfin_new Trello backlog<br/>Redesigned and developed the dashboard web interface of the new project<br/>Discussions with Junior Developer<br/>Discuss with Junior about setting up a Dolfin account<br/>Discuss with Junior about the transition of the project (React part)<br/>Discuss with Junior about the account delete functionality<br/>Discuss with Juniors about the handover document<br/>Connect the reported financial well-being feature to the database and backend(Dolfin_new)<br/>Connect income and expenditure overview to the backend(Dolfin_new)<br/>Connect D-cloud to the backend(Dolfin_new)<br/>Add linking to the bank account feature(Dolfin_new)<br/>Fix the backend Docker file<br/>Complete the showcase video (Dolfin_new part) |
 | Ata Colak |<li>Develop chatbot which uses Groq API as inference engine and LLAMA3-70b as large language model</li><li>Experiment running LLMs locally. Best performing model locally is "Phi3" using "Ollama".</li><li>Introduce logic to pass LLMs only relevant transaction info, reducing risk of hallucination tremendously.</li><li>Introduce capabilities to store different files correlating to different dates</li><li>Got PR merged for the final state of chatbot which extracts date information from user message, searches KnowledgeBase folder for relevant date, and answers user question related to their transactions.</li>
+| Axesh Patel |<li>Solved all the errors and fine-tuned the Llama.cpp model and successfully pushed the code on GitHub.</li><li>Created Dolfin comprehensive FAQ document containing 100 general questions and answers related to customer assistance for the Dolfin application.</li><li>Integrated the FAQ document as a knowledge base for the chatbot, utilizing a Retrieval-Augmented Generation (RAG) model for accurate responses.</li><li>Implemented a new version of the chatbot using Groq cloud's API to leverage Meta's latest LLAMA3 model.</li><li>Enhanced chatbot capabilities by utilizing the newest version of the LLM, improving performance and user interaction.</li><li>Created a detailed report comparing the LLAMA CPP model and the LLAMA Groq API model.</li><li>Added voice recognition and sentiment analysis features to both Llama.cpp and Groq API versions of chatbots.</li><li>Integrated both chatbot models with the front end, providing flexibility to switch between models by simply commenting/uncommenting code.</li><li>Documented the implementation and integration process for both chatbot models, detailing steps.</li>
 | Full name | Write contributions here ! <li>ITEM 1</li><li>ITEM 2</li><li>ITEM 3</li><li>ITEM 4</li><li>ITEM 5</li><li>ITEM 6</li> |
 
 </details>