Published inDev Genius·Jul 1, 2022Distributed Model TrainingDistributed Training Deep learning is a subset of machine learning, a branch of artificial intelligence to perform tasks through experience. Deep learning algorithms are well suited and perform the best with large datasets, not to mention the need for high computation power. With the pay-per-use serverless service model, such as the google…Distributed Systems7 min readDistributed Systems7 min read
Published inDev Genius·Jul 1, 2022Why You Should Reinvent the WheelAs I get slightly older and slightly wiser — to me, it’s more and more apparent that we humans live a life full of contradiction, are rarely self-satisfied, and thrive on learning and discovering new things. …Reinvention4 min readReinvention4 min read
Published inDev Genius·Jun 7, 2022Google Summer of Code — Cartograms in GrapherUpdate: I got the gig 😋 A proposal I submitted for Google Summer of Code, Summer 2022 to Our World in Data (OWID) to implement a population-scaled, shape preserved world cartogram. 1. Overview Our World in Data (OWID) strives to make knowledge of global problems accessible and understandable through interactive geographical data…Google Summer Of Code9 min readGoogle Summer Of Code9 min read
Published inDev Genius·Jun 7, 2022Anatomy of a System Design Interview: The exampleWhat’s better than going over a real system design interview question (Interviewed in June 2022). 1. Requirements Review the end-to-end design below. You need to design Your System. …System Design Interview8 min readSystem Design Interview8 min read
Published inDev Genius·Jun 7, 2022Refactoring NightmaresYou are working on a fast-growing, never-stopping product; 35 sprints and 600 coffees later, you finally got to move on, and a new engineer steps in, or you realize sooner or later that you cannot keep up with the feature requests while ensuring code quality. Then, after raising the need…Refactoring5 min readRefactoring5 min read
Published inDev Genius·May 17, 2022Anomaly Detection and Remediation1. Introduction The overview of the post is on building a system to detect potential anomalies and take immediate action(s). Hence primarily has two phases: Detect and Contain. However, the post discusses the use case to detect and contain anomalies as a generic implementation detail. 1.1. Two main components: Anomaly Detection Service (Detect): Define rules to…Anomaly Detection6 min readAnomaly Detection6 min read
Published inDev Genius·Apr 18, 2022Hybrid Spatial Data Structures: Quad-KD and R-KD trees1. Abstract A hybrid spatial index is a data structure that combines two or more data structures suitable for effectively storing spatial objects to improve search performance. The project report compares tree data structures kd-tree, quad-tree, and r-tree and then introduces hybrid tree structures quad-kd and r-kd trees to reduce the access…Spatial Index10 min readSpatial Index10 min read
Apr 4, 2022Popeye the MightiestWork In Progress. Until then, who do you think is going to win? 🤔Popeye The Sailor2 min readPopeye The Sailor2 min read
Published inDev Genius·Mar 16, 2022Location-based application using Django, PostGIS and LeafletIn a prior post, the experiment on hybrid spatial-index to find spatial data points is worth exploring, but the in-memory implementations and the lack of ability to scale make it almost pointless to use it in real-world applications. In a web application that requires storage and accessing spatial data (co-ordinates)…Postgis3 min readPostgis3 min read
Published inDev Genius·Mar 14, 2022Floating Point Round Off Errors in Geometric AlgorithmsFloating Point Precision in Geometric Algorithms Most geometric algorithms are designed for exact real arithmetic; replacing them with floating-point arithmetic may cause those implementations to fail. Unfortunately, there are no go-to documents on what can go wrong and why across geometric algorithms. …Convex Hull6 min readConvex Hull6 min read
