- Markham, Ontario, Canada - aatifu101@gmail.com AI Dev & Ops at Tasteport (VC-Backed) | Data Science, ML: Python, TypeScript, R, Bash, AWS, Docker • Generative AI & LLMs certified associate [nVIDIA, 2025] I am one of the first 10 early members, focusing on training AI foundational models.
BETA version for Tasteport to fulfill customer needs. Includes vector search - 25+ FAQ and customer support documents for RAG and OpenAI gpt-4o used for generation. Tech stack includes Langchain, Pinecone, OpenAI API keys, FastAPI, Docker, git. Deployed with AWS ECS and EC2 for load balancing for Wix. HTML + CSS for frontend
Invisible City is a geospatial platform that transforms real raw environmental data into an interactive risk-assessment tool. It visualizes facility-level toxicity data and identifies statistical anomalies through ML. Tech Stack: Frontend: React, Vite, Mapbox GL JS, Framer Motion | Backend/ML Pipeline: Python (Pandas, Scikit-learn), ML Models: - Isolation Forest - For unsupervised anomaly detection, Standard & Robust Scalers - For industry-normalized risk scoring
Built with React, NextJS, ThreeJS, TailwindCSS, Git and TypeScript. Deployed with Vercel
Integrated into this website! (bottom right) Tech stack includes Langchain, Pinecone, OpenAI, FastAPI, and Appwrite. Deployed with AWS Lambda
Started position at Tasteport focusing on E-Commerce. Also took first highschool coding course where I was introduced to Python and OOP
Worked with E-Commerce catalog information at grocery store clients. This includes scanning products and organizing database information. Collaborated with ECom Marketing professionals to shoot images, manage SKUs, and organize different product groups. Handled price synchronization and special product categorization.
Handling day to day operations.
Managed live B2B relations with million-dollar business owners and managers. Handled onboarding for inventory tools and operations fulfillment using Java-based tools at grocer client sites. Responsibilities included B2B operations communications with managers, fleet, and fulfillment staff.
Started off in Business and switched into life sciences after first year. Also continued to practice and take programming courses online through MITOpenCourseware(Comp Sci, Maths, Stats), Youtube, Coursera, and LeetCode.
Found lots of interest in the intersection of Computer Science and the Life Sciences.
Started taking data science courses, deep diving documentation and working on learning through projects.
Python - Numpy, Pandas, Matplotlib, sklearn, Pytorch, Huggingface R - tidyverse - Focused on creating projects that could later be applied to either Tasteport (e.g. Chat Assistant, Grocery Classification) - or towards interests in bioinformatics (e.g. Rice Classification) and AI in Healthcare. Also started to work on cleaning, organizing and visualizing datasets in kaggle to understand how to work with real, messy data.
Fine tuning frontier models for grocery product grouping. Also took DeepLearningAI courses such as intro to ML and DL. Transferred to York University.
Contributed to the training and optimization of frontier model for thousands of SKUs for grocery product grouping - Saves hundreds of hours of manual grouping as we scale. Learned how to build ML algorithms and how to build neural networks from scratch + understand the mathematical concepts that enable them.
Recieved Generative AI and LLMs NVIDIA certification. At Tasteport I used this knowledge by building an AI assistant - Tia.
Focused on increasing my knowledge and application skills in AI and machine learning - This certification solidified my ML, DL, Transformers, LLMs, GenAI, and NVIDIA hardware knowledge. See more in the projects section above on Tia.
Created a website which explains and visualizes hidden chemical risk in Toronto through an interactive map and easy to read visuals. Seperately built a similarity-based caching layer.
See Invisible City in project section above and check out my repo on Github (Linked at top of screen)! Also worked on another project where I developed a semantic caching system using PyTorch and Redis to reduce LLM API costs and latency.
I'd love to hear from you. Fill out the form below and I'll get back to you as soon as possible.