Avishek Shrabon

AI Researcher & Interdisciplinary Computing Engineer

+880 1618 569 306 avishek@avishekkerketa.dev avishekkerketa.dev github.com/Avishek7777 linkedin.com/in/avishek-shrabon

Research

HSBN-FL: Hierarchical Stochastic Bottleneck Networks for Heterogeneous Federated Learning

Independent Research — In Progress, Core Experiments Complete

2025 — Present

Addresses a cross-disciplinary challenge at the intersection of distributed systems, privacy, and machine learning: enabling heterogeneous institutions (hospitals, NGOs, research labs) to collaboratively train models without sharing raw data or model weights
Proposes replacing weight exchange with latent-space communication — clients share compressed representations rather than parameters, natively supporting architectural heterogeneity across participants
Server structured as a two-level hierarchy: a Transformer adapter (Z1) attending over all client representations simultaneously, and an MLP apex (Z2) compressing the pooled output into a global abstract state; top-down feedback flows back to each client as a soft alignment signal
Evaluated on CIFAR-100 across 20 heterogeneous clients under Non-IID Dirichlet partitioning; achieved under 0.7% accuracy variance across heterogeneity levels

Hierarchical Stochastic Bottleneck Networks: Conditions for Representational Abstraction in Deep Hierarchies

Manuscript Under External Review — Targeting IEEE Conference

2025

Investigates under what conditions deep networks form a genuine hierarchy of abstractions rather than encoding redundant information across layers — a question spanning information theory, neuroscience, and deep learning
Identifies three necessary mechanisms: per-level objectives, bandwidth-limited stochastic channels (KL penalty), and bidirectional message passing — each level learns a strictly more compressed representation than the one below
Verified empirically on CIFAR-10 and CIFAR-100 using Centered Kernel Alignment and intrinsic dimensionality; ablations isolate two qualitatively distinct failure modes
Manuscript complete preparing for submission to a lab and finally a IEEE conference

Adaptive Federated Learning with Heterogeneous Data and Client Distributions

Undergraduate Thesis — AIUB Academic Repository, Grade: A+

2024 — 2025

Proposed a pipeline combining gradient-based client clustering with cross-architecture Knowledge Distillation (ResNet and ViT) to address non-IID data distributions in federated settings; taught myself the domain independently in approximately six weeks
Supervised by Debajyoti Karmaker, formerly AIUB, now Lecturer at RMIT & AIH, Australia

Education

Bachelor of Science in Computer Science & Engineering

American International University Bangladesh (AIUB) — Dhaka, Bangladesh

Jan 2022 — Dec 2025 CGPA 3.74 / 4.00

Maintained a scholarship continuously from my third semester through graduation.
Awarded the Dean's Award (Spring 2022–23) for academic excellence — CGPA above 3.8
Notable coursework: Computational Statistics & Probability (A+), Algorithms (A), Software Requirements Engineering (A+), Basic Mechanical Engineering (A+)

Experience

Private Tutor — Sciences & Mathematics

Self-employed — Dhaka, Bangladesh

Mid 2022 — Present

Tutored 3–4 students simultaneously across Grade 8–12 in Physics, Chemistry, Biology, Mathematics, Higher Mathematics, and ICT — sustained alongside a full-time degree and independent research
Developed strong ability to adapt explanation strategies to individual learners; regularly communicated complex ideas across different levels of prior knowledge

Projects

1000 Missionary Movement Bangladesh — Training Platform

Full-Stack Web Application — Live Deployment, Nationwide NGO

2025 — Present

Collaborated with NGO leadership and programme coordinators to translate a manual, paper-based intake process into a digital platform — required sustained cross-functional communication with non-technical stakeholders
Built on Next.js, Prisma, and Auth.js; handles applicant intake, role-based access, and programme administration across hundreds of participants nationwide

Church Event Ledger & Ticket Reconciliation System

Full-Stack Software — Live Deployment, 5 Consecutive Events

2025

Identified an operational bottleneck through direct observation of event logistics; built and deployed a working solution in 3 days in coordination with the event team
Reduced end-of-night accounting time from 15–20 minutes to under 5 minutes — a 75% reduction; adopted as standard practice across subsequent events

Proximity Sensor Aid for the Visually Impaired

Hardware — Embedded Systems & Assistive Technology

2023

Bridged electrical engineering and human-centred design to build a wearable obstacle detection system using Arduino Nano, HC-SR04 ultrasonic sensor, and experimental Time-of-Flight sensor; critically evaluated limitations and proposed improvement path

Darkness-Triggered Street Light Automation

Hardware — Electronics & Infrastructure

2023

Designed an LDR-based automatic street lighting system triggering relay-controlled lights based on ambient darkness — motivated by energy efficiency and public infrastructure improvement

Skills

Languages Python, JavaScript, TypeScript, SQL, C++ (surface), C# (surface)

AI / ML PyTorch, scikit-learn, Hugging Face, NumPy, Pandas, Seaborn, Matplotlib

Web & Frameworks React, Next.js, Nest.js, Node.js, Vite, HTML, CSS/SCSS

Databases PostgreSQL, MongoDB, SQLite

Tools Git, GitHub, VS Code, Google Colab, Kaggle, Linux

Research Federated Learning, Knowledge Distillation, Distributed Systems, Information Theory, LLMs, NLP

Awards & Activities

Dean's Award

American International University Bangladesh

Spring 2022–23

Runners-Up — Annual Volleyball Tournament

American International University Bangladesh — competitive team sport

2024

Volunteer — Church Community Services

Event coordination, software deployment, and community engagement — ongoing

Ongoing

References

Anjir Ahmed Chowdhury

PhD Researcher, Intelligent Data and Systems Lab — University of Houston, USA

aachowd4@cougarnet.uh.edu

Debajyoti Karmaker

Lecturer (RMIT & AIH), Australia — Undergraduate Thesis Supervisor

Further contact details available upon request.