Manikandan Ravikiran

About Me

I am Manikandan Ravikiran, working across language, vision, time series, and robotics. Previously, I was a Research Scientist and Lead Researcher at Thoughtworks AI Labs, following nearly a decade at Hitachi R&D. Across these roles, I have been drawn to a common challenge: building models that learn from limited or imperfect information, and understanding when their predictions deserve our trust.

My research centers on representation learning, structured reasoning, and the reliability of AI systems. I have developed methods for understanding multilingual text, assessing educational content, and identifying information lost in generated summaries, alongside work on efficient object recognition, tracking, and multimodal document understanding. More recently, I have been investigating forecasting and anomaly detection, with a particular interest in explaining temporal predictions, as well as robotic perception and planning through haptic sensing, few-shot learning, and language-guided interaction. Across these problems, I aim to develop methods whose behavior can be understood, evaluated rigorously, and translated into practical systems.

Google Scholar · GitHub · Email

Collaborators

Research has been a collaborative effort throughout my career. I have had the opportunity to work with researchers across universities and industrial laboratories, including:

• Rohit Saluja and Arnav Bhavsar — IIT Mandi; my doctoral advisors.
• Martin Klinkigt — Kyocera Innovation Labs, Japan.
• Shinichi Satoh and colleagues — National Institute of Informatics, Japan.
• Andreas Dengel, Sheraz Ahmed, and Jorn Hees — DFKI, Germany.
• Snehanshu Saha — APP CAIR, BITS Pilani.
• Vinay P. Namboodiri — University of Bath (Prev. IIT Kanpur).
• Bharathi Raja Chakravarthi — National University of Ireland, Galway.
• Gaurav Sharma — IIT Kanpur, INRIA-THOTH, and NEC Media Analytics, USA.
• Anand Kumar Madasamy — NIT Karnataka, Surathkal.
• Sangeetha Sivanesan — NIT Tiruchirappalli.
• Ratnavel Rajalakshmi — VIT Vellore.

Selected publications and preprints

A selection of established contributions and recent work. Follow the linked records for publication details and the full author lists.

Language, reasoning, and educational AI

• Unsupervised Informal Language Identification Via Geometric Structural Refinement
  Manikandan Ravikiran, Tanmay Tiwari, Vibhu Gupta, and Rohit Saluja (2026). In Proceedings of the 2026 IEEE International Conference on Systems, Man, and Cybernetics (SMC), Bellevue, Washington, USA, October 4–7, 2026.

• Inference-Time Soft Preference Shaping for Budgeted Abstractive Summarization Manikandan Ravikiran (2026). In Proceedings of the 2026 IEEE International Conference on Systems, Man, and Cybernetics (SMC), Bellevue, Washington, USA, October 4–7, 2026.

• TEEMIL: Towards Educational MCQ Difficulty Estimation in Indic Languages
  Manikandan Ravikiran, Siddharth Vohra, Rajat Verma, Rohit Saluja, Arnav Bhavsar. COLING 2025. Datasets and baselines for estimating question difficulty in Hindi and Kannada.

• DIFFICODE: A Dataset and Graph Spectral Framework for Programming MCQ Difficulty Estimation
  Manikandan Ravikiran, Rakesh Prakash, Arnav Bhavsar, Rohit Saluja. IEEE Access, 2026.

• Hi-SEMFLOW: Lie Algebra–Based Semantic Flow for Span-Level Informal Language Identification in Hindi
  Manikandan Ravikiran, Tanmay Tiwari, Vibhu Gupta, Rohit Saluja. CHiPSAL 2026.

• GISA: Gradual Information Selection Attention for MCQ Difficulty Estimation
  Manikandan Ravikiran, Tarun Sharma, Arnav Bhavsar, and Rohit Saluja (2025). In Artificial Intelligence in Education: 26th International Conference, AIED 2025, Proceedings, Part I, pp. 193–206. Springer.

Vision and multimodal learning

• Multi-Layer Pruning Framework for Compressing Single Shot MultiBox Detector
  Pravendra Singh, Manikandan R, Neeraj Matiyali, Vinay P. Namboodiri. IEEE WACV 2019.

• Hi-GoTE: Hierarchical Group-Wise Temporal Ensembling for Semi-Supervised Pedestrian Attribute Recognition
  Manikandan Ravikiran, Soumen Biswas, Ananth Ganesh. IEEE ICMLA 2023.

• Structural Auxiliary Supervision for OCR of Low-Legibility Handwritten Answers
  Shrey Chandola, Manikandan Ravikiran, Rohit Saluja. Array, 2026.

• DKT: A First Look at Dynamic Kernel Tuning for Pedestrian Attribute Recognition
  Rahul Mishra, Manikandan Ravikiran, Soumen Biswas, and Ananth Ganesh (2024). In Proceedings of the 2024 IEEE International Conference on Big Data (BigData), pp. 5192–5200.

• A Sensitivity Analysis (and Practitioners’ Guide to) of DeepSORT for Low Frame Rate Video
  Manikandan Ravikiran, Yuichi Nonaka, and Nestor Mariyasagayam (2020). In Proceedings of the 2020 IEEE International Conference on Big Data (Big Data), pp. 5227–5236.

Time-series forecasting, anomaly detection, and interpretability

• NVExplain: Explaining Time Series Forecasting with Latent Trajectory Analysis and Structure-Preserving Surrogates
  Muyan Anna Li, Manikandan Ravikiran, and Aditi Gautam (2026). In Proceedings of the 2026 IEEE International Conference on Systems, Man, and Cybernetics (SMC), Bellevue, Washington, USA, October 4–7, 2026. Preprint.

• NVTesseractAD: Dual-Axis Difficulty Scheduling in Conditional Diffusion for Multivariate Time-Series Anomaly Detection
  Aditi Gautam, Manikandan Ravikiran, Saira Qureshi, Sena Fatma Ekiz, Charan Jagwani, Mahmoud Elhadidy, and Sanjeev Roka (2026). In Proceedings of the 2026 IEEE International Conference on Systems, Man, and Cybernetics (SMC), Bellevue, Washington, USA, October 4–7, 2026. Related preprint, under an alternate title.

• Memory-Augmented Forecasting: Scalability and Generalization Across Temporal Domains
  Aditi Gautam, Manikandan Ravikiran, and Fatma Sena Ekiz (2025). In Proceedings of the 2025 IEEE International Conference on Big Data (BigData), pp. 7258–7265.

• Beyond MAE: Measuring Forecast Reliability with Temporal Dependence-Aware Error (TDE)
  Manikandan Ravikiran, Aditi Gautam, and Alisha Chulani (2025). In Proceedings of the 2025 IEEE International Conference on Big Data (BigData), pp. 7271–7277.

• Adaptive Ensemble Imputation for Trip Data using Deep Reinforcement Learning (AEI-DRL)
  Ankit Sharma, Akhash Vellandurai, Thiruvengadam Samon, Vinoth Kumar, and Manikandan Ravikiran (2024). In Proceedings of the 2024 IEEE International Conference on Big Data (BigData), pp. 5211–5219.

Robotics and autonomous systems

• Signal-To-Syntax: A Language-Based Framework for Few-Shot Classification Using Robot Manipulator with Haptic Perception
  Tarun Sharma, Manikandan Ravikiran, Deepika Gupta, Radhe Shyam Sharma, and Rohit Saluja (2026). In 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). Poster.

• Prompted to Fly: Translating Free-Form Instructions into Schema-Constrained Mission Generation for UAVs Using LLMs
  Ankit Sharma, Manikandan Ravikiran, Sohom Chakrabarty, Rohit Saluja, Mrunmayee Limaye, Harshal Kolhe, and Samiron Biswas (2025). In Proceedings of the 2025 IEEE International Conference on Big Data (BigData), pp. 7278–7286.
  Natural-language mission authoring for UAV ground-control systems, including MAVLink and planner-specific mission formats.

More publications on Google Scholar →

Experience

• Thoughtworks AI Labs — Research Scientist / Lead Researcher, April 2025–August 2026
  Led research across NLP, computer vision, multimodal AI, and time-series modeling; worked with researchers and engineering teams to turn methods into prototypes, evaluation systems, and enterprise applications.

• Hitachi R&D — Research Scientist, June 2015–March 2025
  Developed applied AI for mobility, surveillance, healthcare, cybersecurity, and industrial analytics, spanning robust perception, clinical NLP, efficient learning, and temporal modeling.

Education

• Ph.D., Computer Science and Engineering — Indian Institute of Technology, Mandi Advised by Prof. Rohit Saluja and Prof. Arnav Bhavsar (Thesis under review)
• M.S., Computer Science — Georgia Institute of Technology, 2022.
• B.E., Computer Science — Visvesvaraya Technological University, 2015.

Recognition and research community

• Outstanding Reviewer Prize — ACL 2023.
• Best Paper Runner-up — AIED 2025.
• Best Paper — BHASHA Workshop, AACL-IJCNLP 2025.

I contribute to the research community through reviewing, workshop organization, teaching, and mentoring. My service includes ACL, EMNLP, AACL, EACL, COLING, LREC, IEEE SMC, IEEE Bigdata, ACM WSDM, IEEE IROS and several other leading venues. I have helped organize workshops and tutorials on resource-constrained ML, Dravidian language technologies, and cross-modal learning, and Legal AI.

My research has also led to plethora of granted patents covering data-efficient object recognition, heterogeneous distillation, learning from noisy data, and inference scaling.

Contact

For correspondence about my work, please email manikandan.ravikiran@gmail.com.