CV
Resume (PDF)
Research Interests
Affective Computing, Emotion Recognition, Multimodal Large Language Models, Cognitive Science for AI
Education
- Nanyang Technological University, Singapore (Aug 2024 – Present)
- Doctor of Philosophy, Computer Science. Supervisor: Prof. Erik Cambria
- Research on emotion understanding in multimodal LLMs grounded in cognitive appraisal theory, covering future-emotion prediction in dialogue and MLLM evaluation on affective art and narrative interpretation.
- Indian Institute of Technology Roorkee, India (Jul 2016 – Jul 2020)
- Bachelor of Technology, Mechanical Engineering. GPA: 8.04/10
Research Experience
- Artificial Intelligence Lab, NTU Singapore, Research Assistant (Aug 2024 – Present)
- Supervisor: Prof. Erik Cambria
- Developing a model to predict future emotional states, combining evolutionary prompt optimization with reinforcement learning.
- Surveyed cognitive science theories of affect and analyzed how they are applied in current AI research, datasets, and evaluation methods.
- Evaluated large language models (LLMs) and multimodal LLMs (MLLMs) on affective tasks, including student engagement in educational settings, narrative interpretation, and affective art understanding.
- Robert Bosch Centre for Cyber-Physical Systems, IISc Bangalore, Research Assistant (Jun 2022 – Jul 2024)
- Supervisor: Prof. Suresh Sundaram
- Proposed OGP-Net, an RGB–IR fusion method robust to missing modalities, using automatic mixed feature exchange and spectral knowledge distillation for semantic segmentation (AAAI 2025).
- Developed SKD-Net, spectral knowledge distillation for segmentation in low-light thermal imagery (ICRA 2024).
- Implemented a rotation-equivariant detector and Oriented R-CNN to improve small-object detection in remote sensing imagery.
- AI and Robotics Technology Park (ARTPARK), IISc Bangalore (Mar 2021 – May 2022)
- Research Associate (Jun 2021 – May 2022); Research Intern (Mar 2021 – May 2021)
- Supervisors: Prof. Suresh Sundaram, Badrinarayana Rangarajan
- Trained and benchmarked YOLOv3, Faster R-CNN, and SSD for aerial object detection, reaching 93% accuracy.
- Cut YOLOv3 inference latency by 40% (20 ms → 12 ms) on an NVIDIA A6000 via pruning and knowledge distillation.
- Built a gRPC client–server pipeline for bidirectional UAV data streaming over Wi-Fi, removing dependence on AirSim and ROS APIs.
- Simulated single- and multi-UAV missions with integrated object detection (MAVSDK, AirSim), and teleoperated a ground robot using the ROS navigation stack.
Industry Internships
- Robert Bosch Engineering and Business Solutions, Summer Intern (May 2019 – Jul 2019)
- Applied nonlinear Model Predictive Control (VIATOC) to overhead-crane obstacle avoidance and benchmarked auto-generated C++ solvers on a Raspberry Pi 3B.
- iX Energy Pvt. Ltd., Summer Intern (May 2018 – Jul 2018)
- Modelled an electric vehicle’s electrical, mechanical, and cooling systems and wrote Stateflow-based ECU code for start-up, error handling, and CAN communication.
Selected Project
- IIT Roorkee Motorsports (Formula SAE), Vehicle Dynamics and Controls Lead (Jul 2017 – Jun 2020)
- Led the Vehicle Dynamics division for RMS E-19, a formula-style electric race car.
- Built vehicle dynamics models and a lap-time simulator whose design optimizations cut lap time by 2 s.
- Developed SLAM and Model Predictive Control algorithms for a simulated autonomous Formula Student car.
Publications
J. Teotia, X. Wang, X. Zhang, R. Mao, E. Cambria. Evaluating Multimodal Large Language Model Narrative Interpretation through the Lens of Appraisal Theory. LREC 2026, pp. 11417–11426.
X. Wang, J. Teotia, R. Mao, W. K. R. Singh, S. B. Tiun, E. Cambria. Appraisal Theory-Informed Emotion Prediction. LREC 2026, pp. 11345–11358.
A. Sikdar, J. Teotia, S. Sundaram. OGP-Net: Optical Guidance Meets Pixel-Level Contrastive Distillation for Robust Multi-Modal and Missing Modality Segmentation. AAAI 2025, pp. 6922–6930.
J. Teotia, X. Zhang, R. Mao, E. Cambria. Evaluating Vision Language Models in Detecting Learning Engagement. IEEE ICDM Workshops (ICDMW) 2024, pp. 496–502.
A. Sikdar, J. Teotia, S. Sundaram. SKD-Net: Spectral-based Knowledge Distillation in Low-Light Thermal Imagery for Robotic Perception. ICRA 2024.
Technical Skills
- Languages: Python, C++, LaTeX
- ML & LLMs: PyTorch, TensorFlow, scikit-learn, Hugging Face Transformers
- Robotics & Simulation: ROS, PyBullet, AirSim, MATLAB/Simulink, SolidWorks
- Tools: Docker, Git
Awards & Achievements
- 1st in Acceleration and 3rd Overall, Formula Green 2020 (Formula Student competition).
- MCM Scholarship, IIT Roorkee (2 years).
- All India Rank 1986 in JEE Advanced 2016 among 150,000 candidates.
Service & Outreach
- Mentored seven first-year Mechanical and Industrial Engineering students (Student Mentorship Program, IIT Roorkee).
- Taught high school students and organized blood donation camps (National Service Scheme); led a mushroom-farming training initiative under Unnat Bharat Abhiyan.