Ideas & Exploration

Research &
Upcoming Work

Context — My research interests emerged from connecting Computer Architecture coursework to machine learning. Realising that bit-width representation in hardware directly maps to weight quantization in neural networks led me to explore Binary Neural Networks, 1-bit LLMs, and on-device inference pipelines independently of any formal ML course.
Active Research Areas
Post-Quantum Cryptography (PQC)

Exploring next-generation security standards by researching and modeling Post-Quantum Cryptography systems, specifically focusing on lattice-based mathematical frameworks to resist quantum computing attacks.

PQC Lattice-Based Crypto Cybersecurity
Model Quantization (1-bit / 4-bit / 8-bit)

Exploring how reducing the bit-width of neural network weights reduces memory and computation cost — enabling powerful models to run on constrained hardware. The Digit Recognition project served as a testbed for binary input encoding.

BitNet BitNet b1.58 INT4 / INT8
Binary Neural Networks

Studying BinaryConnect and XNOR-Net — networks where weights and activations are constrained to {−1, +1}. XNOR and popcount operations replace multiply-accumulate, making BNNs fast on standard hardware without special accelerators.

BinaryConnect XNOR-Net Binary Activations
1-bit LLMs — BitNet & BitNet b1.58

Studying Microsoft Research's approach to extreme quantization in large language models. BitNet b1.58 constrains every weight to {−1, 0, +1} with comparable performance to full-precision models, making LLM inference viable on commodity hardware.

BitNet Paper b1.58 LLM Inference
On-Device AI Inference & Edge AI

Studying end-to-end deployment pipelines for running quantized models locally on mobile and embedded devices — zero latency, no cloud dependency. Directly motivates the scam detector idea.

TFLite ONNX Runtime On-Device
Upcoming Project Ideas
Concept
AI Scam Call Detector for Low-End Phones

A locally-running AI system to detect scam calls in real-time on low-end Indian smartphones. Uses a quantized speech model to analyse call audio on-device — no cloud, no data sharing, instant alerts.

Edge AI Quantized Speech On-Device Android
Exploring
Full Binary Neural Network from Scratch

Extending the Digit Recognition project to a complete BNN — binary weights, activations, and inputs throughout. Goal: maximum accuracy at minimum memory footprint, deployable on ESP32.

BNN 1-bit Weights XNOR-Net ESP32
Planned
Open Source ML Library Contribution

Following the OpenCV contribution, planning to contribute to a Python ML library — NumPy, PyTorch, or Scikit-learn. Focus: numerical precision, performance optimisation, or quantization-related utilities.

Open Source Python NumPy / PyTorch
Planned
BNN Research Paper

If BNN-from-scratch experiments yield interesting accuracy/memory tradeoff results on constrained devices, documenting findings in a structured technical report or workshop paper.

Research LaTeX Edge Devices