Muhammad M.
AI/ML Engineer
Experience
Dec 2023 - Present
1 year 11 monthsAI/ML Engineer
Software Alliance
- "Just Walk Out" system: integrated YOLOv5 for object detection and segmentation, combined with Deep SORT for real-time person tracking
- Computer vision solutions: deployed real-time tracking using YOLOv8 and object segmentation for accurate product identification
- Sensor integration: implemented weight sensors and virtual polylines to detect and manage item interactions
- Edge computing: utilized edge computing for low-latency processing and real-time decision-making
- User experience enhancement: enabled seamless shopping experience by automating purchase tracking and billing
- Operational efficiency: improved store operations by reducing checkout queues and optimizing resource management
- RAG-based systems: developed retrieval-augmented generation systems using LangChain and Hugging Face, leveraging Meta LLAMA model with Ollama for enhanced language understanding and response generation
Jun 2023 - Jan 2024
8 monthsML Engineer
Kyaas Solutions
- Developed chatbot: created a lead generation chatbot for shoe sales using data scraping, JSON, and PDF files; implemented a knowledge base with Faiss and LangChain for enhanced model flexibility and contextual responses
- Integrated LLM for chatbot: utilized a large language model to generate dynamic responses based on user queries about product details, availability, and sales, including direct product links
- AI-generated video creation: designed a system to generate videos from user-provided topics by creating scripts with GPT, transforming them into scenes, and integrating text-to-speech for narration
- Visual synchronization: synchronized visual elements with voice-over using extracted keywords, ensuring a cohesive and engaging video experience
- Video assembly: merged voice narration with visual scenes and utilized APIs like Pixel API for video delivery, producing informative and visually appealing content
Sep 2022 - Dec 2023
1 year 4 monthsML/NLP Engineer
Conzummate Tech
- Designed AI grammar correction system: developed an end-to-end solution for Swedish grammar correction, achieving ~90% accuracy using supervised learning techniques
- Data collection & preprocessing: leveraged Python libraries such as BeautifulSoup for web scraping, Requests for data retrieval, and regex for text cleansing; utilized Pandas and NumPy for data manipulation and preparation
- Model training & optimization: trained supervised and unsupervised models using Scikit-learn for feature extraction and Keras for deep learning; fine-tuned LSTM models to enhance grammar correction accuracy
- Natural language processing: applied NLTK for text analysis and linguistic data processing, including tokenization, part-of-speech tagging, and parsing
- API development: engineered APIs for grammar correction using Flask and FastAPI, facilitating seamless user interaction and integration with other applications
- Technical proficiency: demonstrated expertise in Python for data science, machine learning, and NLP, leveraging a range of libraries and tools to deliver a high-precision grammar correction system
Dec 2021 - Sep 2022
10 monthsML Engineer
DBG Technologies
- AWS integration: utilized AWS services such as EMR and Glue for building a scalable data processing pipeline; employed S3 for data storage and access
- ETL development: created and executed ETL scripts using Apache Spark on EMR clusters for efficient data preprocessing and cleaning
- Data management: processed and structured large volumes of data, storing the cleaned data back in S3 buckets
- Backend development: developed the backend using Flask, exposing APIs for seamless interaction between the frontend and backend systems
- Network graph creation: employed NetworkX to process telecommunications data and construct network graphs for social network analysis
- Graph visualization: integrated Cytoscape for interactive visualization of network graphs on the frontend, enabling users to explore and analyze graph structures
- Algorithm implementation: applied various graph algorithms, including centrality and betweenness centrality, to uncover patterns and detect criminal networks
Jun 2021 - Dec 2021
7 monthsDS/ML Engineer
SlimLogix
- Dataset & frame extraction: utilized the UCF Crime dataset, extracting frames from videos using annotated files
- Sequence creation: sequenced extracted frames to form video sequences
- Model training: developed a 3D convolutional neural network using TensorFlow to classify crime categories such as abuse, fire, and explosions
- Real-time detection: implemented the 3D CNN model for real-time crime detection, enhancing surveillance and security capabilities
- GAN training: employed conditional generative adversarial networks to generate custom logos
- Conditioning & noise addition: introduced noise and conditioned the GAN on specific logo classes to create user-defined logos
- Logo generation: enabled generation of logos based on user-specified classes and noise inputs
- Custom dataset creation: developed a custom dataset by recording employees speaking each alphabet from A to Z
- Data processing: processed video data of lip movements using 3D convolutional neural networks
- Model training for lip reading: trained the 3D CNN model to recognize spoken alphabet letters from lip movements
- Challenges: addressed challenges such as low video quality and dataset preparation to lay groundwork for future improvements
Summary
With experience in software life cycle processes, I specialize in AI/ML/NLP/CV and Python development. My expertise includes Natural Language Processing, Large Language Models, Computer Vision, and Data Engineering. I have a strong track record of developing and implementing innovative Al solutions, collaborating with cross-functional teams to drive enterprise-level projects.
My passion lies in integrating cutting-edge Explainable and Responsible Al technologies to enhance customer experiences and deliver impactful solutions.
Languages
English
AdvancedPunjabi
AdvancedUrdu
AdvancedEducation
Virtual University of Pakistan
BS · Mathematics · Pakistan
Certifications & licenses
AI Classroom Series
Microsoft
Amazon Web Services (AWS) Complete Data Analyst Bootcamp
Udemy
Introduction To Machine Learning
Microsoft
Machine Learning From Basics
Udemy
Planning A Machine Learning Course To Advance Project
Udemy
Python 101 For Data Science
IBM
Python Programming Complete
Udemy
Need a freelancer? Find your match in seconds.
Try FRATCH GPT More actions
Similar Freelancers
Discover other experts with similar qualifications and experience
- English
- Deutsch
2025 © FRATCH.IO GmbH. All rights reserved.