Experience

• 

  Ptcl.official

  Jul 2019 - Aug 2019

  Intern

  During work at PTCL,i Analyzed the working of Long Communication systems and Learned how the Online complaint system work.

• 

  Al-khawarizmi institute of computer science (kics), uet lahore

  Nov 2021 - Aug 2022

  Research officer (machine learning engineer)

  - Designed and deployed an Urdu OCR System using CRNN (Convolutional Recurrent Neural Network) and LSTM (Long Short-Term Memory) architectures. The system processed input images and accurately extracted Urdu text strings. Integrated YOLO to localize text regions in images, increasing the model's efficiency and accuracy.- Optimized model performance on AWS EC2 instances, demonstrating proficiency with cloud-based machine learning infrastructure.

• 

  Milestonezero

  Mar 2022 - Jun 2023

  Machine learning engineer

  - Implemented a Shoplifting Detection System using YOLOv8 for human detection, DeepSORT for object tracking, and TRTpose for accurate pose estimation. The system operated efficiently on NVIDIA Jetson platforms, with models optimized using TensorRT and Deepstream for real-time performance.- Developed a Content-Based Movie Recommender System using BERT embeddings for natural language processing and cosine similarity for movie content comparison. The system was designed to personalize recommendations based on user preferences, improving the overall user experience.- Built an Image Similarity Model using Vision Transformers (ViT) to detect and analyze image resemblances, which enabled precise comparisons and content filtering for image-based applications. Show less

• 

  Evolve innovative solutions

  Jul 2023 - Jun 2024

  Ai engineer

  - Designed a Leadership Coach Chatbot using the LangChain framework and fine-tuned custom Language Models (LLM) with Qlora for domain-specific responses. Integrated ElasticSearch as the vector database for efficient query handling and fast retrieval of results.- Developed a Content Creation Chatbot using FastAPI and LangChain to automate content generation across 25 categories. Implemented GuardRails to ensure response safety and quality, with real-time checks for content appropriateness. Successfully integrated the chatbot into WhatsApp and Facebook Messenger for broader accessibility and seamless user engagement.- Fine-tuned Stable Diffusion models for e-commerce by creating a customized dataset for face and background transformations on mannequins. Combined this with Segment Anything and Roop models to provide a user-friendly tool for live transformation applications.- Developed a text-to-image search system using embeddings for similarity comparison and ElasticSearch for efficient vector storage, enhancing the search functionality of the platform. Show less

• 

  Northbay solutions

  Jun 2024 - now

  Senior ai engineer

  - Developed a Property Finder Chatbot leveraging AWS Bedrock for scalable AI model deployment. Utilized LangChain for managing complex conversational workflows and maintaining multi-turn context. Integrated Retrieval-Augmented Generation (RAG) for generating detailed property descriptions and summaries. Connected with real estate APIs for real-time property listings, enabling seamless submission of listings by users.- Built an advanced Customer Support Chatbot using LangGraph and knowledge graph integration to handle complex user queries. Implemented RAG to enhance the relevance and accuracy of responses. Designed the backend with PostgreSQL, where users could execute update, delete, and add operations after confirmation.- Employed LangChain and AWS SageMaker for natural language processing, model management, and deployment, making the AI systems scalable and adaptable to changing business needs.- Worked with AWS services including Amazon Q, Polly, Transcribe, Comprehend, SageMaker, Lambda, CloudWatch, S3, and IAM for various AI, automation, and deployment tasks. Show less