Oct 2021 – Aug 2022 · 10 mos · Saarbrücken, Saarland, Germany

• - Wear forecasting of Industrial tools in Glass production ecosystem.

Sep 2020 – Sep 2021 · 1 yr · Saarbrücken, Saarland, Germany

• Developed a machine learning model to predict price of used cars based on various exogenous factors. Implemented an explainable AI module(SHAPly) to provide transparent insights into how different car attributes influence price predictions, enhancing model interpretability and user understanding.
• Conducted a comprehensive user survey to assess the impact of explainable AI on price justification. Results demonstrated that providing users with AI generated explanations improved their acceptance and trust in the predicted prices by 12%

Jul 2021 – May 2022 · 10 mos · Dresden, Saxony, Germany · On-site

• Developed 𝗔𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻-𝗯𝗮𝘀𝗲𝗱 𝗘𝘅𝗽𝗹𝗮𝗶𝗻𝗮𝗯𝗹𝗲 𝗡𝗲𝘂𝗿𝗮𝗹 𝗡𝗲𝘁𝘄𝗼𝗿𝗸 𝗺𝗼𝗱𝗲𝗹𝘀 for machine data analysis and remote sensing. Quicker on-site analysis and faster human interaction to prevent failure were made possible by reducing the number of important frequencies to a few essential frequencies for detecting the anomaly in low-memory rotatory machine systems
• 𝟮.𝟮𝘅 𝗶𝗺𝗽𝗿𝗼𝘃𝗲𝗺𝗲𝗻𝘁 𝗶𝗻 𝗲𝘅𝗽𝗹𝗮𝗶𝗻𝗮𝗯𝗶𝗹𝗶𝘁𝘆 of classification model by introducing a new custom designed loss function over attention maps to obtained concise and differentiable binarized attention maps with increased focus on relevant region with higher confidence and substantial reduction in uncertain regions.

Oct 2018 – Mar 2020 · 1 yr 5 mos · Islamabad

• 𝐃𝐨𝐫𝐦𝐚𝐧𝐜𝐲 𝐏𝐫𝐞𝐝𝐢𝐜𝐭𝐢𝐨𝐧 𝐌𝐨𝐝𝐞𝐥𝐬: The purpose of these models is to predict the customers who will move from 4G to non-data and 4G to 3G data.
• 𝐃𝐞𝐦𝐚𝐧𝐝 𝐅𝐨𝐫𝐞𝐜𝐚𝐬𝐭𝐢𝐧𝐠 𝐌𝐨𝐝𝐞𝐥:
• Demand Forecasting for mobile Recharge and customer Gross Adds keeping in consideration effects of exogeneous variables including promotions, taxes, brand perception etc.
• 𝐀𝐧𝐚𝐥𝐲𝐭𝐢𝐜𝐬 𝐎𝐩𝐬:
• Developed a complete productionization pipeline for data science models to ensure smooth execution of models and marketing campaigns.
• 𝐁𝐞𝐡𝐚𝐯𝐢𝐨𝐫𝐚𝐥 𝐒𝐞𝐠𝐦𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧: Segmentation to split good quality Sims based on their behavior of first 15 days of new sale. This enable business teams to personalize their offers and engage customers accordingly.
• 𝐕𝐨𝐥𝐮𝐦𝐞 𝐅𝐨𝐫𝐞𝐜𝐚𝐬𝐭𝐢𝐧𝐠 𝐌𝐨𝐝𝐞𝐥: Demand Forecasting for monthly data and voice consumed in the next year. Models are developed on cell-site level with separate models for 2G,3G and 4G. This allow technology team to manage their network capacity on each cell site.
• 𝐐𝐮𝐚𝐥𝐢𝐭𝐲 𝐨𝐟 𝐒𝐚𝐥𝐞 𝐏𝐫𝐞𝐝𝐢𝐜𝐭𝐢𝐨𝐧 𝐌𝐨𝐝𝐞𝐥:
• This prediction model predicts bad quality sales based on their usage behavior of first 7 days of sales. Business monitor performance and distribute commission among retailers as measured by this model.
• 𝐂𝐡𝐮𝐫𝐧 𝐏𝐫𝐞𝐝𝐢𝐜𝐭𝐢𝐨𝐧 𝐌𝐨𝐝𝐞𝐥:
• Churn Prediction modelling based on user behavior of last month. Separate models were developed for 2 , 3 and 7 day active base.
• 𝐂𝐚𝐮𝐬𝐚𝐥 𝐀𝐧𝐚𝐥𝐲𝐬𝐢𝐬:
• Performed number of causal & exploratory data analysis to find actionable insights.

Jun 2017 – Oct 2018 · 1 yr 4 mos · Lahore, Pakistan

• 𝐋𝐞𝐝 𝐝𝐞𝐯𝐞𝐥𝐨𝐩𝐦𝐞𝐧𝐭 𝐚𝐧𝐝 𝐝𝐞𝐩𝐥𝐨𝐲𝐦𝐞𝐧𝐭 of core healthcare system modules (outpatient, appointments), providing on-site training to medical staff.
• 𝐏𝐞𝐫𝐟𝐨𝐫𝐦𝐞𝐝 𝐢𝐧-𝐝𝐞𝐩𝐭𝐡 𝐝𝐚𝐭𝐚 𝐚𝐧𝐚𝐥𝐲𝐬𝐢𝐬 of the entire patient journey to identify trends, patterns, and opportunities for process optimization.
• 𝐁𝐮𝐢𝐥𝐭 𝐢𝐧𝐭𝐞𝐫𝐚𝐜𝐭𝐢𝐯𝐞 𝐰𝐞𝐛-𝐛𝐚𝐬𝐞𝐝 𝐫𝐞𝐩𝐨𝐫𝐭𝐬 𝐚𝐧𝐝 𝐝𝐚𝐬𝐡𝐛𝐨𝐚𝐫𝐝 to visualize key performance indicators and drive data-informed decisions
• 𝐂𝐨𝐥𝐥𝐚𝐛𝐨𝐫𝐚𝐭𝐞𝐝 𝐰𝐢𝐭𝐡 𝐬𝐭𝐚𝐤𝐞𝐡𝐨𝐥𝐝𝐞𝐫𝐬 including hospital staff to understand business needs and translate them into actionable data-driven solutions.
• 𝐎𝐩𝐭𝐢𝐦𝐢𝐳𝐞𝐝 𝐬𝐲𝐬𝐭𝐞𝐦 𝐩𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 by implementing code enhancements to improve search and cash payment module efficiency.

Jul 2016 – Jan 2017 · 6 mos

• Computer Architecture and Organization

Jan 2016 – Jun 2016 · 5 mos

• Data Structures and Algorithms

Feb 2015 – Oct 2015 · 8 mos · Islamabad

• - Developed algorithms for extracting desired information from routers output and perform timely and optimized actions based on this information.

Dec 2014 – May 2018 · 3 yrs 5 mos · Islambad