Showcasing Innovation and Success
Explore EnDevSols’ cutting-edge portfolio, where innovation meets practicality. Our expertise includes custom chatbots, deep learning, NLP, computer vision, and custom software development. See how we’ve helped businesses succeed and join us in shaping the future of technology.
SalesAgent – Twillio/GPT
SalesAgent disrupts the marketing landscape by seamlessly integrating GPT-3.5 with real-time phone calls. This innovative project empowers businesses (like BinaryXZ LLC) to leverage the power of AI-generated conversation, delivered with incredibly low latency through Twilio, to create human-like interactions that drive sales.
Finance/Stock Assistant
The Stock Price AI Bot is designed to provide interactive financial analysis and insights. Utilizing Streamlit for web interface development, yfinance for fetching real-time financial data, and LangChain integrated with OpenAI’s GPT-3.5 model for intelligent conversational AI, this project stands at the intersection of finance, data science, and AI.
Law Assistant using GPT 3.5
The Law Assistant app provides a conversational AI interface. The standout feature, ‘Summarize Agreements’, introduces a unique capability to the application. Users can upload agreements or related documents in PDF format, and the application employs advanced text extraction techniques, including OCR to process these documents.
Digitalaalim
Digitalaalim revolutionizes Islamic learning with an intelligent chatbot that serves as your personal scholar on digitalaalim.in. Harnessing the power of GPT-4 and LongTrainer, this chatbot boasts comprehensive knowledge of the Quran, Hadith, and other Islamic scriptures.
Digitalaalim – Admin
Digitalaalim – Admin empowers you to construct a next-generation chatbot, Digitalaalim, by providing a robust suite of information gathering methods. This project equips you with tools for: OCR (Google Vision AI), Whisper AI, Google Translate API etc
Medical LLama 2 Chatbot
Medical LLama 2 Chatbot is trained on a comprehensive medical dataset. This empowers the chatbot to engage in informative and interactive conversations, potentially transforming patient-doctor interactions.
Brainjee Tutor Bots
Brainjee Tutor Bots revolutionize online education with a diverse range of AI-powered tutors designed to cater to your specific learning needs on Brainjee.com. This project offers virtual tutors specializing in various subjects, including English, Math, Science, Psychology, and various languages like Urdu, Hindi, Italian, German, Arabic, and more.
Brainjee Content Generation Pipeline
Brainjee Content Generation Pipeline streamlines educational content creation for Brainjee.com. This innovative project automates syllabus and content generation, crafting personalized learning experiences tailored to specific grades, textbooks, and provided outlines.
Brainjee AI Assistants
Brainjee AI Assistants elevate your experience on Brainjee.com by offering a spectrum of intelligent virtual assistants catering to your personal development needs. This project introduces AI companions specializing in various domains, including Personality Development, Health, Public Speaking, and creative Storytelling.
Gemma LLM Finetuning
This project refines the capabilities of Gemma, a large language model (LLM), to excel at question-answering tasks on the OpenOrca dataset. By leveraging supervised fine-tuning techniques, the project enhances Gemma’s ability to process instructions answer questions To achieve this, the project employs TRL, a powerful method for fine-tuning LLMs.
LLaMa 2 13B Finetuning
This project focuses on refining the capabilities of LLama 2, a large language model (LLM), specifically the 13B parameter version. It aims to enhance LLama 2’s ability to generate typosquat domain names. The project utilizes a self-supervised fine-tuning approach with Transformer Reinforcement Learning on a dataset containing typosquat domains.
TinyLLaMa Finetuning
This project focuses on enhancing the conversational abilities of Tinyllama. Through a self-supervised fine-tuning approach, the project aims to improve Tinyllama’s ability to answer questions and engage in conversations effectively. The project utilizes the OpenOrca question-answering dataset and GPT-4 question formats for training. .
ChatGenius using LLama 2
Unleash the power of AI with ChatGenius, a state-of-the-art, interactive conversational agent powered by the latest large language model (LLM), ‘Meta LLaMa 2 13b chat.’ Designed to simulate human-like conversations and provide contextually accurate responses, ChatGenius redefines the boundaries of AI in conversation and engagement.
Private LLama 2 with LangChain
Are you tired of manually answering frequent questions from your customers? Do you have a large collection of FAQs in PDF format that could be put to better use? Look no further! Our innovative solution utilizes the power of NLP and machine learning to create a chatbot that can answer questions directly from your PDF documents.
AI Summarizer using LLama 2 LLM
The project demonstrates a powerful application of AI in text summarization, overcoming the challenges of lengthy documents by creatively employing the LLama 2 Large Language Models. It not only speeds up the analysis of extensive texts summarization but also maintains the quality and context of the summaries
Logo Generator Stable Diffusion
We’re utilizing the stable diffusion model (CompVis/stable-diffusion-v1-4) from HuggingFace and fine-tuning it for the specific task of logo generation. 🎨💼 Our data muse is the ‘modern-logo-dataset’ provided by Logo Wizard, which boasts a diverse collection of contemporary logo designs.
TrOCR Fine-tuning for Arabic
This project propels deep learning technology into the intricate universe of Arabic diacritics text recognition, an extremely challenging task in the field of NLP. We are utilizing the TrOCR model, employing a Vision Transformer as an encoder and the AraBERT model as a decoder.
Property Report Generation Using OpenAI
We leverage the power of AI to bring insightful property data analysis right to your screen! We’re using real-time property data and the GPT-3.5 to generate comprehensive reports and charts based on several key indicators such as Local Rental Market Analysis, Economic Analysis.
GAN Image-to-Image Translation
Welcome to our project, where we delve into the fascinating world of deep learning to transform grayscale images into vibrant, full-color RGB images! Using the power of Generative Adversarial Networks (GAN), this project focuses on image-to-image translation, a task that involves the conversion of an image from one domain (grayscale) to another (RGB).
Social Event Detection using GNN
In this project, we dive deep into the ocean of deep learning to enhance the results of Knowledge-Preserving Incremental Social Event Detection via Heterogeneous GNNs! We worked with the KPGNN model, along with BERT embeddings and GInConv layers. This fusion of natural language processing and graph neural networks can bring a new depth to social event detection.
Arabic CRNN based OCR
Welcome to our project, where we are pioneering Arabic OCR using deep learning technologies! We are employing a deep Convolutional Recurrent Neural Network and enhancing it with Connectionist Temporal Classification loss. The combined power of CRNN and Connectionist Temporal Classification (CTC) provides a robust way to recognize Arabic text from images, even handling complex Arabic diacritics!
Real-time Market Analysis with Online Learning
Welcome to our cutting-edge project where the worlds of deep learning and finance intersect! In this endeavor, we’re using online learning to make real-time predictions of market exchange rates. The heart of our system is the SGDRegressor model, known for its efficiency in dealing with data that’s constantly changing and growing, perfect for the dynamic world of financial markets.
Q/A Generator Using LMMs
This project showcases the transformative potential of AI in handling and utilizing extensive corporate databases. By employing multiple state-of-the-art language models and customizing them for specific corporate needs, the Q/A generator transforms the way companies access and interact with their data. It not only improves efficiency but also opens up new possibilities for data-driven decision-making and customer engagement.
Bert/Roberta Large Finetuning
Unlock the full potential of pre-trained language models (PLMs) like BERT and RoBERTa with this comprehensive fine-tuning project! We leverage the power of Google Colab Pro to fine-tune large BERT/RoBERTa models on the popular SQuAD question-answering dataset, alongside a diverse range of sequence classification, fill masking, and text generation tasks. Techstack used in TRL, Huggingface, Accelerate and Colab