Federated Learning Leakage Study
A research case study on federated learning privacy leakage, gradient inversion risk, and defense trade-offs for sensitive collaborative training.
PyTorchFederated LearningPrivacyGradient LeakageSecure Aggregation
Resume
Professional profile for AI agents, trustworthy AI, and security-aware technical work.
Jason Chia-Sheng Lin is a doctoral researcher at National Yang Ming Chiao Tung University whose work connects AI agents, trustworthy AI, speech intelligence, cybersecurity, and deployment-ready system design. This page is designed for hiring managers, technical leaders, collaborators, and organizers who want a concise, professional overview.
- Current role
- Doctoral researcher at NYCU
- Base
- Artificial Intelligence in Medical Imaging / Signal Analysis Lab
- Primary focus
- AI agents, speech intelligence, and cybersecurity
- Recent public signals
- 1 official session + 2 English conference papers
Professional Summary
I bring together medical AI lab research, system building, and investigation-informed thinking to design AI agents and IDE-like agent systems that stay useful when evidence, regulation, and deployment constraints matter.
Before doctoral research, Jason worked in cybercrime investigation. That background continues to shape a practical approach to evidence, adversarial behavior, traceability, and the gap between a model that looks good in isolation and a system that remains trustworthy in real-world use.
Current work spans AI agents, speech and language pipelines, trustworthy AI evaluation, medical AI cybersecurity, and deployment-minded system design in environments where reviewability and operational constraints matter.
What Jason Brings to a Team
Agent and systems thinking
Works across models, tools, orchestration, runtime assumptions, and operational constraints rather than treating AI work as an isolated modeling problem.
Security-aware technical judgment
Brings threat modeling, privacy, leakage risk, and deployment realism into system design for environments where failure has real cost.
Clear communication for mixed audiences
Turns technical work into talks, case studies, and structured writing that hiring managers, researchers, and technical collaborators can inspect quickly.
Professional Experience
Current
Doctoral Researcher, NYCU Artificial Intelligence in Medical Imaging / Signal Analysis Lab
Researching trustworthy AI systems, AI agents, medical cybersecurity, speech intelligence, grounded LLM workflows, and security-aware evaluation for real-world deployment.
Previous
Cybercrime Investigation
Worked on digital evidence, online fraud analysis, OSINT, and operational reasoning in high-stakes investigative settings.
Cross-Disciplinary
Investigation-Informed Systems Thinking
Bringing evidence awareness, adversarial thinking, and operational discipline into the way AI agents and systems are designed and evaluated.
Ongoing
Research and Technical Communication
Developing research case studies, technical writing, and speaking material around trustworthy AI, agent systems, speech systems, and deployment risk.
Current Areas of Work
Trustworthy AI agents and systems for operational deployment
IDE-like AI agent systems for research, coding, and analyst workflows
ASR + LLM + RAG pipelines for speech intelligence and evidence-aware analysis
Security, privacy, and evaluation for agentic systems used in high-stakes settings
Human review, traceability, and decision support in analyst-facing workflows
Selected Work
Browse all projects Representative projects that show applied systems thinking.
These case studies show how Jason frames problems, builds systems, and explains technical choices in ways that are inspectable by both technical and cross-functional readers.
Fraud Conversation Analysis with RAG
A research-led case study on retrieval-augmented fraud conversation analysis, designed to keep LLM outputs grounded in transcript evidence for high-stakes review.
PythonRAGLLM PipelinesTranscript AnalysisEvidence Grounding
Speech Evidence Intelligence Pipeline
An evidence-aware speech intelligence pipeline using ASR, retrieval, and LLM extraction to turn long-form conversational audio into structured, reviewable outputs.
PythonASRWhisperRAGLLM Pipelines
Speaking and Publication Signals
CYBERSEC 2026
AI 軟體醫材的資安實戰：從美國 FDA 524B 規範到 Threat Modeling 與 Patch SLA 的完整落地
Breakout session on cybersecurity practice for AI software medical devices, using FDA 524B to connect threat modeling, SBOM, Zero Trust design, and auditable risk governance in heavily regulated environments.
CISC 2025 · Conference Paper
Evolution and Defense Challenges of Ransomware-as-a-Service in the AI Era: A Technical and Strategic Analysis Using Medusa and CrazyHunter as a Case Study
English conference paper examining how AI-era RaaS operations evolve through BYOVD, LOTL, covert C2, and adaptive tradecraft, then mapping those threats to a ZTAID-grounded zero-trust defense strategy.
CISC 2025 · Conference Paper
Integration of Threat Pulse Modeling into the ZTAID Zero Trust Maturity Assessment Model: An Analytical Framework
English conference paper proposing Threat Pulse Modeling (TPM) as a way to translate live cyber threat intelligence into ZTAID maturity signals for continuous zero-trust assessment.
Writing and Technical Communication
essay
From Flat UI to Spatial Interface
Liquid Glass, visionOS, and AI point toward a future where the operating system is redefined less by the kernel than by a new operating surface for search, orchestration, and cross-app work.
essay
Minimal Disclosure for Fraud Intelligence: Cross-Node Pattern Formation in High-Stakes AI
A research-oriented essay on cross-node fraud intelligence, minimal disclosure, and trustworthy AI design for high-stakes pattern formation under fragmented evidence.
Professional signals
- Interdisciplinary profile spanning doctoral research, agent and system building, and investigation-informed reasoning.
- Comfortable in research, engineering-adjacent, and technically cross-functional conversations.
- Public work includes an official CYBERSEC 2026 session and two English CISC 2025 conference papers.
Methods and Technical Toolkit
AI / Agent Systems
PyTorchTransformersAI AgentsLLM PipelinesRAG Systems
Speech / Language
ASRSpeech IntelligenceTranscript ProcessingEvidence ExtractionConversation Analysis
Security / Operations
CybersecurityDigital ForensicsOSINTFraud AnalysisFederated Learning Security
Research / Evaluation
Experiment DesignEvaluation FrameworksReproducible WorkflowsPythonGitHub Actions
Open to the Following Conversations
The strongest fit is with teams, labs, or organizers working near trustworthy AI, speech and language systems, deployment-sensitive workflows, or cybersecurity-minded system design.
Research collaboration and interdisciplinary lab conversations
Speaking invitations for AI agents, speech systems, or cybersecurity
Hiring, technical peer exchange, and IDE-like agent system design
Contact
Email is the best route for hiring conversations, research collaboration, speaking invitations, and technically specific discussion.