Projects across systems, games and machine learning.
Asynchronous multiplayer Battleship where solving Codeforces problems earns you turns. A Rust backend handles the rules and keeps players in sync over WebSockets.
Rust / Axum / Tokio / Next.js
The server owns game rules and coordinates match state. Codeforces submission checks connect programming progress to gameplay, while reconnect handling keeps a match recoverable. Public launch feedback informs iterations on the rules and interface.
An imperfect-information card game where lying and catching lies matter. I’m exploring whether agents can learn an opponent’s habits when they play against the same person over several rounds.
Python / PyTorch / FastAPI / Next.js
Experiments compare agent variants against scripted opponents over seeded matches. The current evaluation is synthetic; human play and broader generalization are not established.
A Pokémon battle-agent project investigating decision-making with hidden opponent information, legal action constraints and reproducible baselines.
Python / PyTorch / Pokémon Showdown
Pinned FoulPlay and Metamon baselines complete legal local Gen 9 OU battles. Paired measurement is in progress.
The current focus is reproducible comparison against pinned third-party baselines. Competitive strength, benchmark-server conformance and public ladder performance have not been evaluated.
An explainable diabetic-retinopathy screening prototype combining retinal image analysis, lesion evidence and Grad-CAM to support clinician review.
MATLAB / Python / PyTorch / Simulink
Strong internal results did not transfer to an external population. Domain shift became the central evaluation lesson.
The current MATLAB pipeline reports sensitivity 0.908 and specificity 0.877 on a sealed APTOS test split (733 images). With its threshold unchanged on Messidor-2, sensitivity falls to 0.363, specificity is 0.865, and AUC is 0.666. These are separate internal and external evaluations, not clinical deployment evidence. The repository is private; this summary is based on the project's current evaluation reports.
Quality gating, vessel/lesion analysis, calibration and Grad-CAM support the workflow. Clinician assessment and clinical validation remain open.
A Rust CLI for todos, with escalating consequences for missed deadlines: notifications, audio and desktop effects. Optional file deletion is disabled by default.
Rust / Linux / systemd
I started coding in eighth grade because I wanted to mess with a friend and turn him into a Discord bot for our server.
Python / discord.py
A few projects I’ve worked on with teams through classes, clubs and hackathons.
Interactive DSA visualizations built through Openverse. Our team gave a lightning talk at OOSC 3.0 and presented the tool at T4E 2025, IIT Madras. Selected for OOSC 4.0.
JavaScript / React / Framer Motion
An online judge built with a team at IIIT Kottayam. I worked on contest lifecycle, staff workflows, submission dispatch validation and per-test verdict persistence.
Next.js / TypeScript / PostgreSQL
A student-status proof-generation prototype using Google JWTs and Noir. On-chain credential verification remains unfinished.
Next.js / Solidity / Noir
Team Hogorithm's trauma-support prototype, awarded Best Software at IEEE MACE .hack(); '26. I contributed Kotlin safety routing, domain models, UI and tests.
Kotlin / Android
A team image-geolocation project using ResNet, quadtree geographic cells and transfer learning. Experiments compare data efficiency and adaptive cell methods.
Python / PyTorch / Computer vision
A completed parallel-computing team project based on Ray Tracing in One Weekend, with serial and OpenMP renderers.
C++ / OpenMP / Computer graphics
I enjoy exploring ideas through implementation and experiments, especially when algorithms, learning and game theory meet.
Competitive programming, data structures and algorithms, and understanding what happens inside systems through C++ and Rust.
Strategic interaction, imperfect information, deadlocks and optimal strategies. I’m interested in how incentives change an agent’s decisions.
Machine learning, deep learning and data science, including recent advances in language models and how to evaluate them.
Understanding model behavior, reliability and alignment, and how to evaluate systems when data, incentives or environments change.
I spent part of my childhood in Hamburg, later lived in Bengaluru, and now study Computer Science and Engineering at IIIT Kottayam.
I’m fluent in English, Kannada, German and Hindi.
Built Edhanta’s education platform, including course progression, parent study summaries, role-based access, authentication and payment flows.
Next.js / React / Better Auth / Drizzle / Neon / Razorpay
Magazine websites, reusable interfaces and editorial tools for publishing and managing content.
Next.js / React / Supabase / PostgreSQL
Algorithm visualizations and open-source work on Algorithima.
Programming-contest organization at IIIT Kottayam.
I compete across several online judges and enjoy the mix of speed, precision and problem solving.
I compete in online CTFs with Team SuKMaDhe. A few of our highlights:
Selected team results. HOLMES details are from the event certificate.
SuKMaDhe on CTFtimeSome of my other interests are fiction, especially Star Wars, The Lord of the Rings and Dune. I’ve read a good chunk of the books too. I also enjoy video games, Pokémon, poker and other card games, basketball, trekking and travel.
I also like ricing my OS (I use Arch btw). I do way more typing tests than a normal person should be doing. Monkeytype
“Specialization is for insects.”
“Do. Or do not. There is no try.”
“Fear is the mind-killer.”
“How much can a crown be worth, when a crow can dine upon a king?”
“Why do we fall, sir? So that we can learn to pick ourselves up.”
“Chief, sometimes you have to jump off a cliff. … Now I ask you, Chief, will you leap with me?”
“Not all those who wander are lost.”
Open to software engineering internships and research opportunities.
prajwal.k141205@gmail.com