Cybersecurity · Data Analytics · Software Engineering · Applied Mathematics

I am Awnon Bhowmik, a doctoral researcher and software engineer working at the intersection of cybersecurity, privacy-preserving machine learning, and applied mathematical modeling. My work combines formal quantitative methods with engineering practice to address security and privacy problems with real deployment implications.
I am pursuing a Doctor of Computer Science, specializing in Cybersecurity and Information Assurance, at Colorado Technical University. My doctoral research sits at the intersection of cybersecurity and privacy-preserving machine learning, with a focus on developing systems that provide formal mathematical privacy guarantees while remaining operationally useful in real security environments.
Professionally, I work as a Computer Systems Analyst and Programmer at the United States Postal Service, where I contribute to the design, development, and modernization of enterprise software systems. Prior roles included software development for financial and healthcare services, and nearly a decade of undergraduate mathematics instruction at CUNY BMCC, where I taught courses from college algebra through differential equations.
My published research spans applied cryptography — including novel encryption schemes, trapdoor function design, and post-quantum approaches — as well as cybersecurity and interdisciplinary work on environmental modeling. I approach research problems with the rigor of a mathematician and the pragmatism of a working engineer, aiming for work that is both theoretically sound and practically relevant.

My doctoral work, conducted as part of the Doctor of Computer Science program (Cybersecurity and Information Assurance) at Colorado Technical University, investigates the application of differential privacy and privacy-preserving machine learning to problems in cybersecurity. The research is grounded in formal mathematical methods and evaluated against practical deployment constraints.
A recurring theme across my research is the tension between formal privacy guarantees and operational utility — understanding when and how theoretical frameworks translate into systems that remain effective under real-world constraints. This line of inquiry draws on tools from probability theory, statistical learning, and algorithm design.
Alongside my dissertation work, I maintain an active program in applied cryptography, with peer-reviewed publications on novel encryption schemes, trapdoor function design, and post-quantum approaches to cryptosystem construction. My published work also extends to mathematical modeling, including interdisciplinary contributions to environmental transport dynamics.
Across all areas, I approach problems with the rigor of a mathematician and the pragmatism of a working engineer. Whether designing adversarially robust privacy-preserving systems or examining how formal frameworks align with regulatory requirements such as GDPR, the throughline is a commitment to work that is both theoretically sound and practically deployable.
Applying differential privacy and privacy-preserving machine learning techniques to cybersecurity problems, balancing formal mathematical guarantees with real-world utility.
Analyzing (ε, δ)-differential privacy mechanisms in the context of classification and anomaly detection tasks. Characterizing the privacy-utility tradeoff across varying privacy budgets and threat models.
Studying the robustness properties of privacy-preserving models under adversarial conditions and their implications for secure system design.
Designing and analyzing novel cryptographic constructions, including hybrid cryptosystems, trapdoor functions, and encryption schemes grounded in number theory and algebraic structures.
Applying analytical and computational models to complex systems — from cryptographic algorithm design to environmental transport dynamics — using formal mathematical frameworks.
Examining how formal privacy frameworks — differential privacy, federated learning — map to regulatory requirements such as GDPR and sector-specific data protection standards.
Research prototypes, published implementations, and engineering work spanning privacy-preserving systems, applied cryptography, and full-stack development.
A browser-based teaching and research tool for differential privacy, featuring real-time mechanism visualization, sequential composition analysis, and a client-side WebAssembly computation engine. Designed for use in lectures, independent study, and LMS integration.
Problem
Differential privacy is mathematically rigorous but often poorly understood in practice. Existing tools either require local installation or lack the interactivity needed to build intuition about privacy budgets, mechanism tradeoffs, and composition behavior.
Approach
Built with Next.js and TypeScript for the interface, with all privacy computations delegated to a WebAssembly module running client-side — no server or database required. Supports Laplace and Gaussian mechanisms with real-time parameter tuning, 13 curated demo presets, sequential composition visualization, and Classroom Pack PDF generation. Includes an embed mode for Canvas, Blackboard, and Moodle integration.
Outcome
Fully functional interactive simulator deployable as a standalone static site. Supports classroom use, independent research exploration, and shareable URL-encoded states for reproducible demonstrations.
A research implementation and preprint introducing a framework for constructing cryptographic elliptic curves using descent methods from arithmetic geometry, producing audit-traceable parameters without reliance on opaque seeds.
Problem
Standard elliptic curve parameter generation often relies on seed-based methods that offer limited transparency into why specific curves were selected, raising concerns about potential backdoor construction. A principled, structure-driven approach with complete audit trails is preferable for high-assurance cryptographic deployments.
Approach
Adapts Selmer's descent theory — using binary quartics and ternary cubics with classical invariants — to deterministically yield candidate curve parameters. Local solubility checks filter candidates before conversion to short-Weierstrass form over prime fields. Each curve is validated against standard cryptographic criteria: group-order analysis, cofactor constraints, twist security, and embedding-degree bounds. The pipeline operates as a Las Vegas algorithm with full audit trail output, and is designed to be compatible with constant-time implementations.
Outcome
Proof-of-concept implementation published as arXiv preprint arXiv:2510.02383 (2025). Companion Jupyter notebook available on GitHub.
A Python-based host and network monitoring system that detects port scans, brute-force attempts, and anomalous traffic in real time, with automated email and SMS alerting, a web-based log interface, and Docker deployment support.
Problem
Many intrusion detection tools are either heavyweight enterprise products or limited proof-of-concept scripts. A self-contained, deployable system bridging research-grade detection logic with operational features — alerting, logging, containerization — is useful for both security research and small-scale deployments.
Approach
Captures live network traffic and monitors process and file system activity for suspicious signatures. Detection heuristics target port scanning patterns and brute-force behavior. Automated alerts are dispatched via SMTP and Twilio SMS. A Flask web interface (port 5000) provides log viewing and event management. Packaged with Docker for portable deployment and includes a PyInstaller build for standalone executable distribution.
Outcome
Fully operational IDS with containerized deployment, automated alerting pipeline, web-based monitoring interface, and unit test coverage.
A host-based security monitoring tool that detects potential keylogging activity through real-time surveillance of processes, sensitive file system directories, and network connections, with configurable alerting and structured event logging.
Problem
Keyloggers operate silently by design and are difficult to detect with conventional antivirus heuristics alone. A behavioral monitoring approach — watching for suspicious process signatures, unauthorized file access, and anomalous outbound connections — provides complementary detection coverage.
Approach
Monitors running processes for known keylogger behavioral signatures, watches configurable sensitive directories for unauthorized access patterns, and flags unusual outbound network traffic. Alerting is handled via SMTP email and Twilio SMS. Events are written to a structured log file for post-incident analysis. Configurable through `config.py`; deployable via Docker or as a standalone PyInstaller executable.
Outcome
Deployable host-based detection tool with configurable monitoring scope, multi-channel alerting, and comprehensive audit logging.
A Python implementation of Shor's quantum factoring algorithm, demonstrating the theoretical mechanism by which sufficiently capable quantum computers can break RSA encryption by efficiently factoring large composite integers.
Problem
RSA security rests on the computational intractability of integer factorization for classical computers. Shor's algorithm achieves polynomial-time factoring on a quantum computer, making it a critical reference point for understanding post-quantum cryptography and the urgency of lattice-based and other quantum-resistant alternatives.
Approach
Implements the quantum period-finding subroutine and classical post-processing steps of Shor's algorithm in Python, illustrating how the algorithm reduces integer factorization to order-finding in a cyclic group and applies the quantum Fourier transform to extract the period efficiently.
Outcome
Working Python implementation demonstrating the factoring pipeline and its implications for classical public-key cryptosystems. Relevant context for post-quantum cryptography research.
This website — a statically generated academic portfolio built with Next.js and TypeScript, featuring a blog system with MDX, LaTeX/KaTeX rendering, and a secure contact form.
Problem
Academic researchers need a professional web presence serving both scholarly audiences (publications, research, CV) and general visitors, without relying on institutional infrastructure.
Approach
Built with Next.js (static export), TypeScript, and Tailwind CSS. The blog system uses MDX with KaTeX for mathematical notation. The contact form includes rate limiting, HTML sanitization, and EmailJS integration. Deployed to GitHub Pages.
Outcome
Production site serving as primary academic and professional web presence. Source code open on GitHub.
Specialization: Cybersecurity and Information Assurance
Colorado Technical University
Mar 2025 – Feb 2030 (expected)
United States Postal Service
Jan 2024 – Present
Data Analytics and Cybersecurity
Colorado State University Global Campus
Aug 2023 – Mar 2025
United States Postal Service
Jun 2023 – Dec 2023
Mathematics and Computer Science
CUNY York College
Jan 2018 – Dec 2019
SS&C Technologies
Mar 2017 – Mar 2023
Mathematics
CUNY Borough of Manhattan Community College
Jan 2014 – Aug 2015
NYC Department of Education
Sep 2020 – Mar 2023
Did not complete
University of Dhaka, Bangladesh
2009 – 2012
CUNY Borough of Manhattan Community College
2014 – 2021
Public contributions to mathematics and computer science education across online scholarly communities.