Exploring the intersection of human and artificial intelligence
Electronics & Biomedical Engineering student at MEC, Kochi. Dedicated to advancing neural decoding, computational neuroscience, and AI-driven medical systems.
import jax.numpy as jnp
from mne.preprocessing import ICA
from tensorflow.keras import layers
@jax.jit
def preprocess_signals(signals):
# Apply high-pass filter and ICA artifact removal
ica = ICA(n_components=20, random_state=42)
return ica.fit_and_clean(signals)
def eegnet_classify(eeg_data):
# Deep CNN for 4-class motor imagery decoding
model = Sequential([
layers.Conv2D(8, (1, 64), padding='same'),
layers.BatchNormalization(),
layers.DepthwiseConv2D((64, 1), depth_multiplier=2),
layers.Activation('elu'),
layers.Dense(4, activation='softmax')
])
return model
I am pursuing my B.Tech in Electronics and Biomedical Engineering at Government Model Engineering College, Thrikkakara (Sept 2024 to Present).
My core scientific interest is bridging the gap between biological brains and computational intelligence. I focus on developing algorithms for Brain-Computer Interfaces (BCI) and extracting meaningful features from noisy neural data.
Driven by curiosity and an entrepreneurial mindset, I enjoy exploring startup ideas, hardware prototyping, and building open-source projects. I actively build my foundations in engineering mathematics, machine learning frameworks, and embedded engineering.
Research internships, open-source leadership, and engineering roles.
Leading workshops, mentoring students, and driving open-source culture at Model Engineering College.
Worked at Tencent AI Lab on ML, NLP, and LLM research for open-source AI model development.
Developed the website for a decentralized marketplace built on the TON blockchain.
Open-source contributions and hardware prototypes combining digital signal processing, Edge AI, and web technology.
Engineered an end-to-end pipeline to classify 4-class motor imagery (hand/foot movement intents) using the EEGNet architecture. Implemented MNE-Python for ICA-based artifact removal, establishing critical performance baselines in BCI experiments.
Deployed a real-time computer vision detection system on an ESP32-S3 Sense board. Optimized a quantized MobileNet V2 model using FOMO algorithm, reaching 7 FPS at 143ms latency for industrial parts sorting.
Contributing to an open-source initiative dedicated to preserving and revitalizing endangered Indian languages. Combining NLP, speech recognition, and LLMs to create accessible linguistic analysis tools.
A high-performance, client-side tool for removing sparkles watermarks from images generated by Google Gemini AI. Uses a mathematically precise Reverse Alpha Blending algorithm to restore original pixels with zero quality loss.
Research preprints and academic contributions in computational neuroscience and AI.
Abhin Krishna · Computational Neuroscience · Neural Modelling
A rigorous comparative analysis of two foundational computational neuroscience models: the biophysically detailed Hodgkin-Huxley model and the computationally efficient Izhikevich model, evaluated for their accuracy in predicting high-fidelity neural spike trains across varied stimulation conditions.
Selected milestones demonstrating academic rigour, research, and global engagement.
Selected for the prestigious Harvard Project for Asian and International Relations Conference.
Completed computational systems neural circuit modeling training at CNS Lab, IIT Madras.
Presented "PhytoScan", an AI-driven fruit freshness detection device at Money Conclave.
Goethe Zertifikat B1 certified speaker in German language with a score of 82.
I am actively looking for co-founders, research collaborators, and start-up connections. If you want to discuss BCI, neural systems modeling, or start-up ideas, let's talk.