I'm a Ph.D. researcher in Natural Language Processing at the UKP Lab, TU Darmstadt, working under the supervision of Prof. Dr. Iryna Gurevych. My research focuses on building safe, explainable, and trustworthy AI systems, with particular emphasis on:

🔬 LLM Safety & Adversarial Robustness - Designing jailbreak attacks and defenses to improve model safety
🎯 Hallucination Detection & Mitigation - Reducing errors in long-form question answering systems
🚀 Production-Ready AI - Shipping scalable NLP infrastructure and low-latency AI services
📊 Model Explainability - Making AI decisions transparent and interpretable

• 📝 2 papers accepted to ACL and EMNLP 2025 on hallucination detection and LLM jailbreaking
• 🏆 Co-led development of UKP-SQuARE, a QA platform used by 1000+ users
• 🛡️ Designed POATE, a contrastive reasoning-based jailbreak attack achieving 40% higher success rates
• 🤖 Built DocChat, a multi-agent RAG system with hybrid retrieval and verification agents

Languages & Frameworks

Tools & Infrastructure

Specializations

• 🤖 LLMs: GPT-4, LLaMA, T5, BART
• 🔗 RAG Systems: LangChain, LangGraph, LlamaIndex, MCP
• 📊 Experiment Tracking: Weights & Biases, LangSmith
• 🔍 Evaluation: Adversarial Testing, Explainability, Calibration

Ph.D. Student @ UKP Lab, TU Darmstadt (Sep 2021 - Present)
Working on LLM safety, hallucination detection, and explainable AI

ML Engineer Intern @ Convaise (Feb 2021 - Jun 2021)
Built automated ML platform reducing deployment time from 2 days to 10 minutes

Research Assistant @ RWTH Aachen University (May 2018 - Apr 2020)
Analyzed 80M+ Amazon reviews for gender bias detection using deep learning

Systems Engineer @ Infosys Limited (Jun 2015 - Aug 2017)
Automated testing and CI/CD, increasing deployment frequency by 400%

• 🔍 Reviewer for ACL Rolling Review (ARR)
• 👨‍🏫 Teaching Assistant for NLP Ethics course (100+ students)
• 🎓 Mentor to 13+ BSc/MSc students at TU Darmstadt
• 🌟 Open to collaborations on trustworthy AI and NLP safety research

• 📧 Email: [email protected]
• 💼 LinkedIn: rachneetsachdeva
• 🌐 Portfolio: rachneet.github.io
• 📱 Phone: +49 176 47194617