Shailja Thakur

Research Scientist IBM Research

News

2026
Two papers accepted at ACL 2026 — main conference and Findings
2026
BenchDrift covered in the press — Crypto Briefing, daily.dev, and edgeX
2026
Invited speaker at Open Source Summit India 2026 (Linux Foundation), Mumbai
2026
Serving on Program Committee for AAAI 2026, EMNLP 2026, COLM 2026
2025
VeriGen received the Best Paper Award from ACM TODAES
2025
Serving on Program Committee for AAAI 2025, AI for Social Impact Track
2024
Serving on Program Committee for WiML 2024, EMNLP 2024 (MRL Workshop)
2024
Our paper VeriGen accepted in ACM TODAES
2023
Two papers on bug fixing and assertion generation using LLMs accepted in IEEE TIFS
2023
Nominated for Best Paper Award — Benchmarking LLMs for Verilog RTL generation, DATE 2023
2023
Invited Reviewer for WiCV at ICCV 2023, AAAI 2023
2022
Post-Doctoral Fellowship in the NSF National AI Institute for Edge Computing (Athena)

Selected talks

Jun 2026
Confidently Wrong: When AI Cannot Catch Its Own Bugs — Open Source Summit India, Linux Foundation, Mumbai
2025
GHCI 2025 — Grace Hopper Celebration India
2024
GHCI 2024 — Innovations and Challenges of AI in Cyber-Physical Systems, Emerging Technologies Track
2024
IBM-IISc AI Day — Data Processing Pipeline/Toolkits from IBM for Gen AI Models Development, Bangalore
Sep 2024
IIIT Hyderabad — Large Language Models for Hardware Design Automation
Aug 2024
PALS Industry-Assisted Lecture Series — Building Successful LLM Applications: Power of High-Quality Data
2024
MSRIT Invited Talk — Large Language Models for Code Generation, Reasoning and Preference Alignment
Feb–Sep 2024
CDS@IISc, IIT Kanpur, IISER Pune, IIIT Delhi, IIT Mandi — Secure, Interpretable, and Scalable ML for Cyber-Physical Systems
2023
Intel, AI and Security CoP — Automating Verilog RTL Code Generation with LLMs, New York
2023
MLCAD 2023 — Tutorial: Intersection of HW, Security & Large Language Models, Snowbird, Utah
Jan 2022
CEDAR, INRIA & Ecole Polytechnique — Security and Interpretability in Automotive Systems
Nov 2021
Duke University, CPS Lab — Security and Interpretability in Automotive Systems
Oct 2021
Arizona State University — Security and Interpretability in Automotive Systems

In the media

Aug 2026
IBM introduces BenchDrift to quantify wording effects in LLM benchmarks — Crypto Briefing
Aug 2026
Benchmark leaderboards might just be measuring phrasing luck — daily.dev
Aug 2026
IBM Research Releases BenchDrift, an Open-Source Tool for Measuring How Prompt Wording Shifts LLM Benchmark Scores — edgeX
2025
AI creates a specialized chip design language, earning top marks — Tech Xplore
2025
NYU Tandon Engineers Create First AI Model for Specialized Chip Design Language, Earning Top Marks — NYU Tandon

Service

2023–26
Program Committee, AAAI 2023, 2024, 2025, 2026
2023–26
Program Committee, EMNLP 2023, 2024, 2025, 2026
2026
Program Committee, COLM 2026
2023–25
Invited Reviewer, NeurIPS 2023, 2024, 2025
2026
Invited Reviewer, IEEE TCAD 2026
2024–25
Invited Reviewer, ACM TODAES 2024, 2025
2023
Invited Reviewer, WiCV at ICCV 2023
2020–
Member, IEEE 2020–Present