Portfolio

Research Profile

Studying how AI changes software work, from the inside of shipped systems.

Prospective PhD applicant working at the intersection of AI-assisted software engineering, trustworthy AI, and human-centered developer tools — combining production AI systems, empirical studies, and machine-generated content detection.

Active Survey My CV Google Scholar
ICPC mock contest — focused team debugging at a lab workstation
ICPC Mock Contest

Competitive programming roots

Origin

Before research papers, there were contest floors and late-night debugging sessions.

Programming contests shaped how I think about problem decomposition, evaluation under pressure, and collaborative reasoning — instincts that now drive empirical questions about software engineering, Software security and Artificial Intelligence.

ProgrammingContest experience
8+Years engineering
3Research papers

Publications

Research works worth opening first.

Two active preprints on trustworthy detection of machine-generated content, plus the IEEE publication that started the research path.

All on Scholar
01
Under reviewCode detectionarXiv preprint · 2026

Whitespaces Don't Lie: Feature-Driven and Embedding-Based Approaches for Detecting Machine-Generated Code

Combines interpretable handcrafted code features with CodeBERT embeddings to distinguish human-written and AI-generated source code.

Read
02
Under reviewImage detectionarXiv preprint · 2026

Handcrafted Feature Fusion for Reliable Detection of AI-Generated Images

Evaluates HOG, LBP, DCT, GLCM, and wavelet descriptors with feature-fusion pipelines to improve interpretable image-generation detection.

Read
03
PublishedNLP foundationsIEEE ICCIT · 2017

Question Classification Using Support Vector Machine with Hybrid Feature Extraction Method

Undergraduate thesis and first publication on NLP question classification using hybrid feature extraction and support vector machines.

Read

Research Direction

Questions I want to keep pursuing in a PhD.

Research that stays grounded in how software is actually built, reviewed, debugged, deployed, and trusted.

  1. 01

    AI-Assisted Software Engineering

    How coding assistants change day-to-day engineering work, where they improve flow, and where they introduce reliability, review, and trust issues.

  2. 02

    Human-Centered Developer Tools

    Interface design, cognitive load, developer confidence, and how tools should fit real software teams instead of demos.

  3. 03

    Trustworthy Detection of Machine-Generated Content

    Practical methods for identifying AI-generated code and images using handcrafted features, embeddings, and interpretable models.

  4. 04

    Secure, Reliable AI Systems

    Dependable deployment patterns for AI systems, including secure pipelines, grounded outputs, and resilient engineering practices.

Current Work

Engineering practice that feeds the research agenda.

Production AI delivery, a live empirical study, and active detection preprints — each reinforcing the others.

Present

AI + Full-Stack Engineering

Designing production AI systems and developer-facing platforms, with direct exposure to reliability, usability, and scaling problems that shape the research agenda.

View Resume

2026 · Ongoing

Empirical study on AI coding tools

Surveying software engineers in Bangladesh on productivity, confidence, collaboration, skill development, and career perception.

Take the Survey

2026 · Under review

Machine-generated content detection

Interpretable features, embeddings, and robust evaluation for detecting AI-generated source code and images.

Google Scholar

Applied Base

Systems that motivate the research questions.

Concrete products and pipelines — not filler portfolio pieces — that expose the gaps between demos and trustworthy deployed AI.

Production AI

Re:tune AI Chatbot Platform

Knowledge-grounded chatbot infrastructure with ingestion, embeddings, semantic retrieval, and response-generation flows.

Data + Retrieval

Data Harmonization Dashboard

CSV/Excel upload, schema mapping, harmonized review, and Pinecone indexing with OpenAI embeddings.

Applied Product

10xlaunch

Visitor identification and lead enrichment into Slack. Product Hunt Product of the Day and Week.

ML Foundations

Language Identification from Text

Multinomial Naive Bayes and Logistic Regression over character and word n-grams from the DSL dataset.

Healthcare AI

iHealthScreen

Diabetic-screening software and retinal-image workflows during an internship in applied healthcare AI.

Developer Experience

Replo Visual Builder

Reusable interface systems and drag-and-drop editing for a visual web-building product.

Industry Signal

Grounded in real software environments.

Access to realistic workflows, product constraints, and trust failures that often disappear in small lab prototypes.

2024 — Present

Re:cruit

Senior Full Stack Software Engineer

Production work across AI, frontend architecture, cloud applications, and high-stakes product workflows.

2022 — 2024

re:tune

Founding Engineer

Built an AI chatbot platform around retrieval, embeddings, and user trust, eventually supporting 20,000+ users.

2021 — 2022

Replo

Software Engineer

Reusable UI systems and drag-and-drop interfaces that sharpened interest in human-centered developer tools.

Collaboration

Open to PhD supervision, research collaboration, and funding conversations.

If you work on software engineering, HCI, trustworthy AI, or developer tools, I would be glad to connect.

Email Me My CV