Masood Ahmed

After getting promoted from associate data scientist to data scientist, i moved first into generative computer vision work and now more into product ownership. right now i’m mainly responsible for the data science side of two products, from planning to production. 1 - Work with generative computer vision models like stable diffusion, flux, etc. to train custom styles and people-specific models (data prep, fine-tuning, evaluation and deployment). 2 - Design and own the data + ml side of two products end-to-end: problem framing, defining metrics, planning experiments, training models and iterating based on results. 3 - Build and maintain pipelines for collecting, cleaning and structuring data used for both traditional ml and deep learning / generative models. 4 - Set up evaluation frameworks (offline metrics, comparisons between model versions, sanity checks) to understand how models behave before and after going live. 5 - Work closely with engineering to integrate models into production, handle inference issues and monitor how models perform on real users over time.

At Walee, i work across data pipelines, deep learning models and product-facing ml systems. most of my work is around taking raw user + product data, turning it into usable signals, and then building and evaluating algorithms on top of it. Here is what i did in details: 1 - Designed and maintained data pipelines to collect, clean and structure data from different sources (product events, user behaviour, platform logs) so it could be used for model training and analysis 2 - Worked with core deep learning algorithms and traditional ml models, including setting up training loops, loss functions, evaluation pipelines and monitoring to see how models behaved on real data 3 - Built a credit scoring model for a lending-style product to estimate how likely a user is to repay a loan, using behaviour, profile and historical patterns as features, and iterating on the model based on evaluation results 4 - Helped design and implement the initial recommendation engine for a short-video platform, combining user behaviour features (what users watch, interact with, skip, etc.) with deep learning models to rank and recommend content 5 - Worked on algorithm evaluation, comparing different model versions and setups using offline metrics and validation strategies, and feeding those results back into how we engineered features and tuned models 6 - Built a text-to-mongodb system, where i fine-tuned multiple LLMs and chained them so a user could type a natural language query, have it converted into an accurate mongodb query, run against the db, and then get the results analysed + summarised back in plain language

I worked with sveston watches on a project basis while i was at walee, mainly helping them on the data side. 1 - Took their raw data (sales, campaigns, etc.) and cleaned / organised it so it was easier to work with. 2 - Did analysis on performance and behaviour patterns to help them understand what was actually happening in their data 3 - built scripts / small pipelines to automate some of the repetitive data work instead of doing everything manually.

At techlets, i mainly worked on the scraping + data side of things. most of my time went into building crawlers, managing data pipelines and writing scripts to move data from messy websites into something clean and usable for the team. Here is what i did in details: 1 - built web crawlers in python using scrapy, selenium, APIs and sitemap-based scraping to collect data from news and ecommerce sites. 2 - handled real-world scraping issues like bot detection, sessions, rate limiting and keeping crawlers stable over time 3 - set up data pipelines to clean / process scraped data and insert it into mongodb 4 -worked on core scripting + logic tasks (small tools, automation scripts, data utilities) to support other parts of the workflow

I did my schooling through home schooling using khan academy, up to the equivalent of 12th grade. I covered the compulsory subjects but spent most of my time on maths and physics, which i enjoyed the most. alongside that, i was learning programming, building small projects and experimenting with things like robotics and basic algorithms.