Oct 2022 – Jul 2025 · 2 yrs 9 mos · London

• Develop systems powering the Bloomberg Terminal, delivering real-time data and insights to 300,000+ users.
• Oversee services handling 400,000+ daily requests for BloombergNEF users, ensuring reliability & performance.
• Consistently optimize code efficiency, including an 87% reduction in memory usage from 10.7 GB to 1.4 GB.
• Collaborate cross-functionally to design scalable systems, integrating queues & caches to enhance performance.
• Manage 12+ Terminal applications, powered by 30+ Python microservices and 10+ JavaScript applications.
• Engineered a client-facing Terminal application to efficiently aggregate asset data from multiple sources.
• Designed and inner-sourced services from scratch, enabling BloombergNEF trials for event attendees globally.
• Crafted an application in the Bloomberg Professional app as part of innovation; later adopted by business.
• Build monitoring dashboards to provide actionable insights into service performance, reducing debugging time.
• Co-lead the Python Guild, fostering 4,000+ internal members to align Bloomberg with the broader ecosystem.

May 2021 – Jul 2021 · 2 mos

• Worked on the Cult.Live platform, which lets us experience live fitness lessons from our home
• Engineered an Android product where users experience live fitness lessons virtually, track their workouts, and challenge others
• The reps of an exercise are computed via computer vision & pose estimation
• Devised a similarity algorithm to calculate the similarity between user’s pose & reference pose and get the score; in both Python & React Native
• Cases such as the user being farther/nearer to the camera and time delays between user & trainer are also handled

May 2020 – Jul 2020 · 2 mos · Doha, Qatar

• I had worked on advanced data analysis & visualizations of indicators that characterize the quality of news articles.
• As a result, I was awarded the 1st place project award among 50+ candidates & 35 projects.
• I had analyzed & interpreted the information, and then turned it into meaningful representations with intricate Business Intelligence tools.
• Some analysis indicators included propaganda content distribution, fake news, centrality, bias, and political ideology.
• I had developed 40+ cross-linked visualizations forming 6 robust dashboards.
• Some visualizations include SPLOM (Scatterplot Matrix), Stream graphs, Marimekko charts, Word Clouds, and Tree/Heat Maps.
• I had worked with the Social Computing directorate and on the Tanbih news platform.
• The Tanbih team plans to integrate my work to the Tanbih platform.

May 2019 – Jun 2019 · 1 mo · Doha, Qatar

• I had worked under the project titled "Deep Learning Architecture for Imputing Gene Data" alongside another fellow intern.
• We were awarded the second-best project amongst approximately 30 projects.
• In this project, we exploited Deep Learning technique to predict and impute the missing values in the SNPs - Genome Population Dataset.
• We tested 2 models: Denoising Autoencoders using Partial Loss (DAPL) & Neural Matrix Completion (NMC).
• We also implemented a fully responsive GUI using the Django framework.
• This experiment can have applications in the diagnosis and treatment of cancer & other diseases of similar genomic profile.