Skill set
Israr in a few words
Hello! I'm a dedicated Machine Learning Engineer with 8+ years of experience in building real-world AI solutions.
I specialize in:
✅ Computer Vision
✅ Deep Learning (CNNs, RNNs, Transformers)
✅ Image Processing
✅ Natural Language Processing (NLP)
✅ Large Language Models (LLMs)
✅ Chatbot Development
✅ Web Scraping
✅ Python Automation
I’ve worked with startups, researchers, and enterprises to bring AI-powered ideas to life. My goal is to not only deliver quality work but also to understand your business needs and help your product succeed.
Whether you're looking to:
- Develop a custom AI model,
- Automate repetitive tasks using Python,
- Analyze images or text with machine learning,
- Build an intelligent chatbot,
- Or scrape data for business insights — I'm here to help.
Feel free to reach out to me if you have an idea, a project in mind, or just need expert guidance.
Let’s turn your vision into reality!
Experience
- Become Current PBCComputer Vision EngineerSeptember 2021 - Today (3 years and 9 months)Scituate, MA, USA• Researched & Developed the complete cloud-based pipeline for the training of Computer Vision models on ec2.• Cloud-trained models are converted into an offline mode to use mobile computation. It helped us save more than $1000 per month for 10,000 users only.• My model optimisation/quantization strategy helped us to improve the processing speed from 3 images to 10 images per second.
- CHINIKI GUARD LIMITEDComputer Vision LeadJune 2019 - August 2021 (2 years and 2 months)London, UK• Real-time updates on the activities inside a retail store, reports shoplifting by analyzing the feed from your CCTV.• Performed market research with team members to identify the technical difficulties, and took steps to make strategies and the product's technical roadmap.• Designed the whole data collection centre to replicate the real world scenarios. Implemented the whole data collection pipeline to make the process easier for the data engineering team.• Led thorough research on related academic papers to understand how our problem can be solved using deep learning, Shortlisted the best methods to help implement our own custom model architecture.• Used Amazon sagemaker for training pipelines using a custom model support feature. Sagemaker ground truth service being used for training dataset creation and storage.• Optimized models for FP16 mode using TensorRT, improved fps from 16 to 23 on Jetson Nx. NVIDIA's DeepStream SDK is used for multi-camera processing.• Used Sagemaker Edge Manager Agent to send data on an S3 bucket and the data is then used for improvement and retraining of the model.
- UpworkMachine Learning EngineerJune 2017 - June 2019 (2 years)• Created customized applications to solve critical problems using machine and deep learning frameworks.• Prototyped machine learning applications and quickly determined application viability.
Recommendations
Education
- Bachelor of ScienceNational University of Sciences and Technology2018BS Computer Science