This is an advertisement for Prof. Gail Kaiser's Spring 2027 offering of COMS E6156 Topics in Software Engineering. 6156 is the only regularly offered course at Columbia CS where students have the opportunity to explore in depth a specific topic of their own choosing. 6156 is a seminar oriented towards graduate students and advanced undergraduates who aspire to become, or already are, researchers or industry technology leaders.  Students should be highly self-motivated to pursue their own forward-looking topic within software engineering, broadly construed. 6156 is not "more" 4156, and not "more advanced" 4156. 4156 is about doing software engineering, and 6156 is about improving software engineering including software security.  Students choose their own forward-looking topics, with instructor approval.  Many students chose to investigate AI4SE or SE4AI.  Most other recent students have studied static/dynamic program analysis, either source level or binary/bytecode, and/or software testing to detect logic bugs, performance bugs, security vulnerabilities, privacy breaches, or malware.  There are no lectures after the first week or two; students present research papers that are then discussed by the whole class. Active participation is required, and students are expected to read the papers scheduled for presentation before attending class. A tentative list of papers is provided on the syllabus, and students can suggest others.  Besides readings, presentations and discussions, the workload consists of a midterm paper and final project. Midterm papers review recent literature in the student's chosen topic. Final projects build an innovative tool or demonstrate an innovative technique. Both midterm papers and final projects should include empirical experiments. Students may optionally organize teams (of any reasonable size) for the final project but midterm papers should be written individually. 6156 is an secondary elective for MS Software Systems and is also accepted for MS Computer Security for students who choose security or privacy topics.  6156 may be accepted as a secondary elective for other MS pathways whose students choose relevant topics (get approval from your advisor **before** the end of the program change period!) and is a technical elective for all CS and CE graduate and undergraduate students. Students should have already completed at least one graduate level software systems course, preferably 4113 (distributed systems), 4115 (compilers), 4118 (operating systems), 4156 (software engineering), 4181 (security) or their equivalent. Some project titles from last year: NeatAgent: A Local AI Agent for Autonomous File Organization Noise categorization and post filtering evaluation for AI Code Review TrustSQL: A Reliability-Oriented LLM-based Text-to-SQL System Behavioral Determinants of Deployed AI Agents in Social Networks: A Multi-Factor Study of Personality, Model, and Guardrail Specification FinVerify: Improving the Reliability of LLM-Based Financial Advisors Through Hybrid Verification Systems SEAL: Software Engineering Agentic Looper for Push-Button Construction of an Operating System You Can't Leak What You Don't Know: Information-Isolated Disclosure for LLM Dialogue Agents Some suggested student topics for this spring: Hybrid Local/Cloud LLM Architectures for Privacy TOR Routing to LLM APIs for Privacy Security Benchmarking for LLM-generated Systems Software CWE-based Security Auditing for LLM-generated Software OS Technologies for Agent Containment Screen/Play Writing as a Prompt Engineering Paradigm Record/Replay Tracing for LLM Bug Repair Metamorphic Testing for Quantum Computing If you would like to take the course, please complete "homework zero", posted at http://www.cs.columbia.edu/~kaiser/6156-choosing-from-waitlist.Spring2027.txt, and email to kaiser+6156@cs.columbia.edu and/or submit in Vergil.