2013 - 2017
Bachelor of Technology (BTech) in Computer Science and Engineering at VIT University, Vellore, India
PCgaming benchmarking enthusiast: overclocking gpus/cpus/memoryClocks, worrying about fan-temperature-voltage control ..Upgrading every generation..pci gen3 to gen5,gtx 9400 to 4090s etc
Obsessed with game benchmarking and performance optimization:
- Meticulously analyzed frame rates, Obsessed on 1% (Percentile , p95,p99,p99.9,p99.99) lows for smooth gameplay..Looking back MFUs might be a better measure of performance than gpu Utilizations and frametime (alteast for ML workloads: theoretical max FLOPS vs actual FLOPs util)
- Experimented with various graphics settings to find optimal balance of visuals and performance
- Occasionally rage quit when encountering frustrating frame drops or stutters/artifacts
Hail PCMR! (eyes always on top left corner of the screen tracking fps drops..focus shifted from videos games to harnessing the power of sand)
Course Work:
- Differential and Difference Equations
- Discrete Mathematical Structures
- Multivariable Calculus and Differential Equations
- Numerical Analysis
- Applied Probability, Statistics and Reliability
- Linear Algebra
- Graph Theory and its Applications
- Computer Programming and Problem Solving
- Object Oriented Paradigm and Programming
- Data Structures and Algorithms
- Theory of Computation
- Algorithm Design And Analysis
- Agent Based Intelligent Systems
- Digital Logic
- Computer Architecture and Organization
- Microprocessor and Interfacing
- Modern Physics
- Electronics
- Basic Electrical and Electronics Engineering
- Mathematical Economics
- Operations Research
- Green Computing
- Software Project Management
Mostly only showed interest and scored well in courses(with some logic involved) as they were easy to compress (the patterns)
Humanities subjects(wordcel courses) took my own sweet time and was on bad end of the bell curve (due to relative grading)
Few Projects:
-
Big Data Analysis Project: Analyzed household electricity consumption data using Hadoop's ecosystem, including HDFS, MapReduce, and Sqoop to handle large datasets. Migrated data from traditional MySQL to Hadoop and used Hive for querying. The project focused on clustering and analyzing power consumption data to extract meaningful insights.
-
Data Analytics Web Application (Datagroker Internship): Developed a web application using Python's Flask framework in a Vagrant environment to dynamically analyze data from AWS S3 via the Boto API. The app displayed word embeddings stored in AWS S3, with Vagrant providing a lightweight development environment. Used Ngrok as a reverse proxy to make the local web server accessible remotely for testing and demo purposes.
-
Simple 3D Animation and VFX Project: Utilized Adobe Premiere to create a dynamic 3D animation with VFX elements, focusing on enhancing bike ride sequences. Implemented advanced techniques to seamlessly blend 3D animated elements with live-action footage. Added realistic fire and smoke effects to intensify scenes and create visually captivating moments.
-
Arduino Irrigation System: Built an automatic irrigation system for indoor plants using Arduino and C++. The system utilized a soil moisture sensor to monitor moisture levels in the soil, triggering a DC motor to pump water when levels dropped below a set threshold. An LCD screen displayed the soil status (Dry, Moist, or Soggy) based on voltage readings from the sensor.
-
Android Mobile Apps: Developed several Android applications using Java and Android Studio. Projects included:
- A simple calculator app with basic arithmetic operations
- A to-do list app with local storage functionality
- A basic weather app that fetched data from a public API
-
Visual Studio Apps for Windows: Created multiple Windows desktop applications using C# and .NET Framework in Visual Studio. Notable projects included:
- A simple text editor with basic file operations (open, save, edit)
- A basic image viewer with zoom and rotate functionalities
- A student management system with CRUD operations and local database integration
-
Custom Linux Kernel Module for Syscall Table Injection: Developed a kernel module that performed syscall table injection to replace system call handlers, allowing for monitoring and modification of kernel behavior. The module utilized CR0 register manipulation to bypass memory protection, a technique that was effective in older kernels. This approach, once feasible in pre-2.6.x kernels, became obsolete in kernel 4.x and beyond due to the implementation of advanced security measures like KASLR (Kernel Address Space Layout Randomization) and KPTI (Kernel Page Table Isolation).
GPUs ain't cheap... cranking up the settings to ultra resolutions is like burning through FLOPs. Suddenly, I'm back to being GPU broke and in need of a job just to keep the pixels/frames flowing at 144HZ refresh rates!
Back to Timeline