Trustworthy AI for software engineering
How can AI systems used by developers expose their assumptions, behave more predictably, and remain understandable enough to support real engineering workflows?
Research profile
Trustworthy AI systems, secure model ecosystems, and empirical software engineering.
My research sits at the intersection of AI-enabled developer tooling, software security, and empirical software engineering. I am especially interested in how model platforms, agents, and developer-facing AI systems fail in real settings, and how we can design better technical and process-level safeguards.
5
Research Tracks
2
Active Projects
1
Submitted Work
3
Collaborations
Research agenda
The same themes introduced on the homepage, expanded here as the lens through which I choose projects and papers.
Security and reliability of LLM-based systems
I am interested in the failure modes created when agentic systems, model hubs, and tool-integrated LLM applications meet unsafe defaults or weak contracts.
Empirical software engineering
Repository mining, developer discussion analysis, and deployment-aware evidence gathering are central to how I frame and validate the research questions.
Flagship paper
The best current example of the kind of research problem I want to pursue in graduate school.
Submitted to TOSEM 2026
An Empirical Study on Remote Code Execution in ML Model Hosting Ecosystems
Cross-platform study of ~45,000 repositories across five ML platforms (Hugging Face, ModelScope, OpenCSG, OpenMMLab, PyTorch Hub) with co-authors Mohammad Latif Siddiq and Joanna C. Santos. Detected security issues using static analyzers (Bandit, CodeQL, Semgrep) and YARA malware signatures: found CWE-502 (unsafe deserialization) in 74.54% and CWE-95 (eval injection) in 15.02% of affected repositories; 10.41% of Hugging Face repos contain security smells. Analyzed 600+ developer discussions to build a taxonomy of security misconceptions; found 6.6% SafeTensors adoption and heavy trust_remote_code usage. Submitted to TOSEM 2026.
Current research track record
Each project is presented with the same amount of information so the page reads as a coherent profile instead of a mixed archive.
In Progress
Endpoint-black-box auditing framework for aligned backdoors in LLM agents using counterfactual environments and discrete-choice estimation.
Targeting PVLDB
Extended SchemaAgent with LangGraph StateGraph architecture and 3-tier auto-repair system, reducing redundant LLM calls by 80%.
Research track
Sentiment Analysis of Anonymous Crisis Reports in Bangladesh
Developed uReporter -- Bangladesh's first anonymous reporting system during 2024 national crisis, analyzing 124 crowd-sourced reports using transformer models.
Research track
Patient-Centric Blockchain Framework for EHR Management
Undergraduate thesis on blockchain framework for EHR management with encrypted off-chain IPFS storage and on-chain Ethereum access control.