Saket Kumar Mall

Adding Spatio-Temporal Dataset Support in Data Catalog

Tue, 26 May 2026 00:00:00 +0530

The problem

GRASS stores two kinds of thing in two different places. Ordinary maps (rasters, vectors, 3D rasters) live in the spatial database. Space-time datasets (STRDS for raster series, STVDS for vector series, STR3DS for 3D raster series) live in a separate temporal database, and are managed by the t.* family of modules.

Recoup AI Agent

Thu, 03 Sep 2026 00:00:00 +0530

Subscription businesses lose money in three ways that look different and are really the same: a card fails, a checkout is abandoned, an invoice goes past due. Recoup works all three through one loop. It watches for revenue at risk, works out why each payment failed, prices the cheapest intervention likely to work, executes it inside hard compliance and stopping rules, and then measures what that was actually worth.

There is no language model anywhere in it. Message copy comes from reviewed templates. The intelligence is in the decision policy and an uplift model, not in generation, and a given seed replays byte for byte.

ChronoFlow

Sun, 01 Jun 2025 00:00:00 +0530

ChronoFlow covers the scheduling side of a team’s week: meetings, deadlines, leave requests, and the approval chain that sits behind all three. I wrote both halves of it, the Spring Boot service and the browser client, over about two months in 2025.

How it fits together

The backend exposes a REST API per workflow, each one backed by its own tables in a MySQL schema I normalized up front. Scheduling data duplicates itself easily once you allow recurring meetings and shared attendees, and the schema was the place to stop that happening.

Real-Time YouTube Sentiment Pipeline

Fri, 01 Aug 2025 00:00:00 +0530

Comment sentiment goes stale fast. A job that runs overnight tells you how an audience felt yesterday, which is no use if you wanted to catch a reaction turning while the video was still climbing. So this one streams.

The path a comment takes

A Cloud Function pulls comments and drops them onto Pub/Sub, which absorbs the bursts that arrive whenever a large channel posts. Dataflow reads from there, classifies each comment, and writes the result out. End to end it stays under a minute.

ATS Resume Ranking Engine

Wed, 01 Oct 2025 00:00:00 +0530

Paste in a job description, upload a stack of resumes, get them back in order. The interesting question is what “in order” should mean.

How it scores

Keyword matching answers that question badly. A candidate who writes “built streaming pipelines on GCP” scores nothing against a posting that asks for “Dataflow experience”, even though they are describing the same work. Term overlap measures vocabulary, and vocabulary is the thing a good candidate is least likely to share with whoever wrote the posting.

My GSoC 2026 Journey with GRASS

Fri, 01 May 2026 00:00:00 +0530

Welcome to my Google Summer of Code 2026 experience blog. I documented my progress here week by week while contributing to GRASS under NumFOCUS.

About the Ecosystem

What is GSoC? Google Summer of Code is a global program that pairs students with open-source organizations to work for 8-12 weeks on a programming projects, helping students gain real-world experience while improving open-source software.
What is GRASS? Geographic Resources Analysis Support System (GRASS) is a powerful, free, and open-source Geographic Information System used for geospatial data management, image processing, and spatial modeling.
What is NumFOCUS? NumFOCUS is a non-profit organization that supports open-source scientific computing, providing fiscal, legal, and administrative support to projects like GRASS.

My Project Overview

Title: Add Spatio-Temporal Datasets Support to GRASS Data Catalog GUI

About

Mon, 01 Jan 0001 00:00:00 +0000

I’m a software developer from India. I finished a Bachelor of Computer Applications at SRM Institute of Science & Technology in June 2026 with a CGPA of 9.42, and I’m now on the Master of Computer Applications there.

The work I have enjoyed most tends to arrive after a feature already works. On ChronoFlow that meant tearing up an indexing scheme that looked fine right up until the schedule views went under load. On the sentiment pipeline it meant working out what a dashboard should honestly show when a stage is degraded and the numbers are only half there.