Portfolio · 2026

Adriteyo Das

About

Heylo, I'm Adriteyo !

I study Computer and Communication Engineering at MIT Manipal, with most of my work revolving around low-level systems, DFIR, and AI.

I like figuring out how things work and occasionally building things that seemed like a good idea at the time.

I also used to play CTFs for Cryptonite .

Outside of computers, football takes up some of my time, along with an unreasonable amount of LARPing in RDR2, Hollow Knight, and Crusader Kings III.

This is my portfolio and blog, where I share my work, things I'm building, and whatever else I find interesting.

Experience

Software Engineer Program (SEP) Intern
June 2026 – August 2026
JPMorgan Chase & Co. · Corporate Technology (Firmwide Planning & Analysis) · Bengaluru, India
- Worked as a Data Infrastructure Engineer within Corporate Technology (Firmwide Planning & Analysis).
- Engineered distributed data pipelines and analytics systems utilizing Apache Spark and AWS.
- Automated cloud infrastructure provisioning and CI/CD pipelines with Terraform, Jenkins, Spinnaker, and ADK.
Research Assistant
Aug 2026 – Present
Carnegie Mellon University · Xu Lab
- Working in Embodied AI under Dr. Xiaolong Wang at Xu Lab.
Head of Research & Development
Apr 2024 – Aug 2026
Cryptonite · MIT Manipal's Official Cybersecurity Team
- Leading R&D initiatives for one of the best academic CTF teams in Asia
- Specialize in Digital Forensics, Incident Response (DFIR), and AI Security challenges
- Developed and deployed challenges for niteCTF 2024 and 2025 — 3rd highest rated Indian CTF globally
- Mentoring junior members in forensics, research, ML/DL, and malware analysis
- Part of core research team exploring intersection of AI and cybersecurity
Research Intern
Jan 2025 – Aug 2025
Indian Statistical Institute, Kolkata · CVPR Research Group
- Worked under Dr Umapada Pal with collaboration from University of Salford and Chinese Institute of Information Science
- Implemented novel Deepfake Detection algorithm that outperformed SOTA methods
- Paper accepted at ACPR-2025 (Asian Conference on Pattern Recognition)

Publications

Enhanced Stress Detection via Intermediate Fusion of Intermodal Signals
Jan 2026
Intelligence-Based Medicine
Elsevier ↗
Non-correlative Feature Interaction for Classification of Real and Fake Scene Images
Nov 2025
8th Asian Conference on Pattern Recognition (ACPR), Gold Coast
Springer ↗
Comparative Analysis of Multimodal Architectures for Effective Skin Lesion Detection
Aug 2025
Frontiers in Artificial Intelligence
Frontiers ↗

Education

Manipal Institute of Technology
2023 – 2027
B.Tech — Computer and Communication Engineering
- CGPA: 9.3
- Top 0.01 percentile in the ICT Department
DPS Ruby Park
2021 – 2023
Class 12 — PCMCS Stream
- 95.2% CBSE · House Prefect · Tech Club Core

Skills

Languages

AssemblyScript
C
C++
C#
Embedded C
Python
JavaScript
Java
Rust
Go

Web & Frontend

HTML5
CSS3
SASS
React
NodeJS
Vite

AI/ML & CV

PyTorch
TensorFlow
OpenCV
CUDA
scikit-learn
XGBoost
GANs
CNNs
Transformers
YOLO

DFIR & Security

YARA
IDA Pro
Ghidra
Volatility
Wireshark
Autopsy
FTK Imager
Sleuth Kit
Binary Ninja

Databases & Big Data

Apache Spark
MySQL
MongoDB
Elasticsearch
Redis

DevOps & SRE

AWS
Docker
Kubernetes
Terraform
Jenkins
Spinnaker
ADK
Cloud Foundry
Ansible
Puppet

Observability

Prometheus
Grafana
Splunk
Jaeger
ELK Stack

Tools & Testing

Git
GitHub
Kafka
Selenium
pytest

Achievements

Rank 01

EnigmaXplore 2024 IIIT Naya Raipur
BITSCTF 2024 BITS Goa
CruXipher 2024 BITS Hyderabad
KashiCTF 2025 IIT BHU Varanasi

Rank 02

TexSAW CTF 2025 UT Dallas
BITSCTF 2025 BITS Goa
CapturePoint CTF 2025 IIIT Vadodara
ExploitX CTF 2025 Top Individual

Rank 03

Codefest CTF 2025 IIT BHU
eHax CTF 2025 DTU Delhi
FooBar CTF 2025 NIT Durgapur

Academic

Achiever's Scholarship MIT Manipal
Finalist JPMorgan Code for Good
Paper accepted ACPR 2025

Volunteering

Head of Deep Learning The Data Alchemists · Official Data Science Club, MIT Manipal
Head of Cybersecurity Research Society Manipal · MIT Manipal
Member Rotaract Club of Manipal · Community Service

Projects

ParaOxidizer

github ↗crates.io ↗

Standalone Rust-native post-training quantizer & inference engine for LLMs without external Python/CUDA dependencies. Converts SafeTensors/GGUF to zero-copy memory-mapped .pox format with Apple Silicon Metal compute shaders, second-order AWQ/GPTQ calibration, INT4/INT8 SIMD, and Ed25519 cryptographic verification.

Raya

github ↗crates.io ↗

Rust-native static malware detection and binary pattern-matching engine. High-throughput multi-pattern Aho-Corasick scanning, recursive Pratt AST rule parser, deep PE32/PE32+ & ELF inspection, x86/x64 instruction disassembly (iced-x86), and Shannon entropy analysis with structured evidence tracing.

FeTorrent

github ↗

High-performance asynchronous BitTorrent engine in Rust with a distributed stack: headless daemon, REST/WebSocket API layer, and CLI/Web dashboard controllers.

MemHawk

github ↗

AI-powered memory forensics platform integrating Volatility 3 with local LLMs (Ollama) for automated triage, anomaly detection, and interactive dump analysis.

Irones

github ↗

Cycle-accurate Nintendo Entertainment System (NES) emulator in Rust with 6502 CPU emulation, PPU rendering, and SDL2 audiovisual output.

Xerxes

github ↗crates.io ↗

High-performance filesystem navigation engine in Rust for macOS & Linux. Provides instant directory jumping via fuzzy pattern matching, a frecency-based ranking algorithm, interactive selection, custom path aliases, and automated shell hooks.

rpsmon

github ↗crates.io ↗

Fast Rust TUI system monitor for macOS & Linux with live process insights, tree navigation, lock selection tracking, watchlists, signal dispatch, and configurable alerts.

ChiefWarden

github ↗

Hybrid malware detection framework — Portable Executable (PE) static analysis combined with gradient boosting ML (98.89% accuracy).

BlueShift

github ↗

Native macOS dual-pane SSH and remote file manager built in Swift and SwiftUI with differential delta transfers (rsync), keychain credential store, and multi-hop bastion jump host support.

Astraeus

github ↗

Real-time orbital collision risk prediction with interactive 3D WebGL visualization and machine learning ensemble classifier (94.2% accuracy).

PassRanger

github ↗

Android password manager in Java with encrypted local storage, optional Firebase sync, Autofill integration, and password entropy analysis using rule-based + ML scoring with HIBP k-anonymity checks.

DropLAN

github ↗

Real-time local network file sharing service with zero-configuration peer discovery, streaming transfers, and QR connectivity.

SimuTrade

github ↗live ↗

Real-time stock trading simulator and portfolio tracker with Prophet time-series price forecasting and technical charts.

GAN_RepairNet

github ↗

Deep convolutional generative adversarial network with U-Net architecture for reconstructing severely corrupted dashcam images.

evil-qwen

github ↗

Interpretability research steering Qwen3.5-4B's behavior by editing just two neurons' activations, run locally on Apple Silicon.

PowerWarden

github ↗

Per-app CPU and energy budget enforcer for macOS.

Open Source

z3 Microsoft Research SMT Solver

Z3Prover/z3 · +16 −0 · PR #9500 ↗

Fixed FPA/BV model-validation soundness bug for Array + Datatype theories (Fixes #9488). Aligned model generation logic and synchronized Floating-Point values with bit-blasted Bit-Vector representatives to prevent split-brain validation failures.

CppInterOp CERN & LLVM Research

compiler-research/CppInterOp · Clang AST & Build Infrastructure

Contributed core Clang AST processing improvements, unified symbol resolution, and macOS test invocation fixes to the C++ interoperability engine used at CERN.

PR #842 ↗ PR #846 ↗ PR #829 ↗

macos-UnifiedLogs

Mandiant (Google) · +36 −6 · PR #97 ↗

Fixed out-of-bounds array access vulnerability in message.rs. Added proper bounds checking for unknown_uuid_index, preventing crash risk on corrupted DSC files.

capa

mandiant/capa · +25 −2 · PR #3176 ↗

Fixed bug in capa's PE feature extraction where module names for extensionless DLL imports (e.g. ntdll) were stripped by rpartition(), causing qualified import rule matching to silently fail. Added regression test suite.

superhuman

Google DeepMind · PR #5 ↗

Built a Python library for IMO Bench providing a clean API with type-safe dataclasses, filtering, and 35 passing tests.

campus-dining

aaditagrawal/campus-dining · +643 −50 · PR #3 ↗

Implemented a comprehensive localStorage favorites/bookmarking system for saving content across the Campus Directory.