Anvi Wadhwa
Hi,
I'm Anvi,
CS undergraduate at VIT Chennai.
Welcome to my website. I'd like to give you a neat, well-defined answer to who I am, but I don't really have one. The only thing I'm reasonably certain about is that I enjoy an unhealthy amount of larping. So, for now, let me just be Anvi Wadhwa, and welcome to Larpland.
Most of my curiosity currently lives in machine learning, particularly in areas involving physical and space sciences, probabilistic modeling, uncertainty quantification, and high-dimensional systems. Lately, I've been spending most of my time going deeper into reinforcement learning and trying to understand the mathematics behind how agents learn, make decisions, and optimize behavior. Partly because it's interesting. Partly because humans don't seem particularly good at it either.
Outside of computer science, I'm into philosophy, business, psychology, literature, and art. I love writing and chasing ideas that push boundaries. I appreciate anything that makes me question the obvious and look at the world sideways.
I don't have a grand philosophy beyond staying curious. There are too many interesting things to learn, too many ideas to explore, and too many questions worth asking. Anyway, that's everything I currently know about myself, and now you know it too.
Congratulations, Now you're about as me as I am.
Algorithmic Trading with Reinforcement Learning
RL-based trading system for Indian equities using PPO and custom environments. Sharpe ratio, Calmar ratio, and max drawdown analysis across a full backtesting suite.
Feature Adaptation for Emotion Recognition in Speech
Multi-branch 1D-CNN fusion using MFCC, Mel spectrogram, and Chroma features on RAVDESS. 5-fold stratified cross-validation for robust evaluation.
Systematic Intraday Trading via GARCH Forecasts
Rolling GARCH volatility forecasts paired with 5-minute market indicators. Systematic signal generation and end-to-end backtesting pipeline.
AI Physiotherapy & Rehabilitation System Ongoing
Real-time lower-limb movement analysis for remote physiotherapy. Pose estimation via MediaPipe and OpenCV, hip, knee, and ankle tracking from live webcam streams.
Sex Work, Machines, and the Fear of Becoming Unnecessary
On labour, intimacy, artificial companions, and the fear that machines may unsettle identity as much as work.
Can we Just Stick to a Tesseract?
A dive into higher-dimensional geometry and why our brains resist it so stubbornly.
God Serves Man, but He is Not a Servant
On the tension between utility and dignity.
The Philosophical and Political Dimensions of Punishment
Punishment as deterrence, retribution, rehabilitation: three philosophies, three different societies.
The Bayes' Awakening
On how Bayesian reasoning quietly reshapes the way we update beliefs under uncertainty.
A Man's Need For Fiction Triumphs Over The Evidence Of Absurdity
On absurdity, suffering, religion, and the human instinct to invent meaning when the universe offers none.
pls spell vheck me
A compact walkthrough of edit distance, Levenshtein distance, dynamic programming, and spell checkers.
The Adventures of Schrodinger's Cat
A long-form exploration of superposition, observation, entanglement, and the strange philosophical life of the famous cat.
Computer Vision Research Intern
Indian Institute of Technology (BHU) Varanasi
Lower-Limb Pose Estimation for Physiotherapy Monitoring
- Developed a real-time lower-limb tracking system using MediaPipe and OpenCV.
- Implemented pose estimation pipelines to detect and track hip, knee, and ankle landmarks from webcam video streams.
- Built joint-angle estimation modules to compute knee and hip flexion angles during physiotherapy exercises using geometric analysis.
- Improved robustness across varying lighting conditions, occlusions, and camera viewpoints for reliable real-time tracking.
- Generated motion logs and CSV-based datasets for movement analysis and future model evaluation.
- Visualized joint trajectories and movement patterns to assist physiotherapy assessment and rehabilitation monitoring.
- Assisted in organizing and preprocessing exercise datasets collected from multiple users for training and testing workflows.
R&D and Curation Lead
Business Innovation Community (BIC), VIT Chennai
- Led research initiatives on emerging trends in technology, startups, AI, and business innovation.
- Curated technical and business-focused content for community events, workshops, and social media outreach.
- Managed content strategy and branding across Instagram and LinkedIn to improve engagement and visibility.
- Coordinated cross-functional teams for smooth execution of events, collaborations, and operational activities.
- Contributed to ideation and planning for innovation-driven initiatives within the community.
Core Committee Member & Evaluator
DEFY Hackathons (2024-2025)
- Evaluated startup, Web3, and technology-focused hackathon projects on innovation, scalability, and product-market fit.
- Mentored participants on pitching, business strategy, and solution validation.
- Designed judging rubrics and assisted in streamlining evaluation workflows.
- Managed event operations, participant onboarding, outreach, and sponsorship coordination.
- Collaborated with organizing teams to execute large-scale hackathon activities and engagement initiatives.