I am Lead Research Scientist at ThoughtWorks Global AI Research, where I focus on enhancing Large Language Model (LLM) adoption through evaluation frameworks targeting completeness, explanation fairness, and decoding optimization. Previously, I was a Researcher at Hitachi R&D India, where I worked on AI for Advanced Driver Assistance Systems (ADAS), with emphasis on reasoning systems, project delivery, and scalable AI pipelines. My expertise spans large-scale data analysis, project management, and the development of mathematically grounded AI tools.

I enjoy generating novel ideas and devising scalable solutions to complex real-world problems. Parallel to my industry work, I am a Ph.D. Student at IIT Mandi, advised by Prof. Rohit Saluja and Prof. Arnav Bhavsar where I study rare and impactful problems in AI with applications to the Education domain. My research interests span across fairness in large language models (LLMs), explanation sufficiency, multilingual and educational NLP, legal AI, low-resource modeling, and deep model compression. I’m especially focused on developing evaluation methods and scalable attention mechanisms for complex reasoning tasks in real-world AI systems.


πŸ”” News

  • April 17, 2025 β€” Delivered an invited guest lecture on Multimodal Generative AI and Theoretical Diffusion Modeling at VIT Chennai.
  • March 31, 2025 β€” Joined ThoughtWorks Global AI Research as Lead Research Scientist.
  • March 20, 2025 β€” Congratulations to Hariharan (NIT-K) on the successful submission of his Ph.D. thesis.
  • March 17, 2025 β€” Delivered an invited lecture on Advanced Reinforcement Learning at NIT Surathkal β€” Thank you to the Department of IT, NIT-K.
  • Nov 20, 2024 - Delivered Keynote on Industrial Computer Vision at IEEE ICDDS 2025
  • Nov 10, 2024 β€” Three papers accepted at IEEE BigData 2024 β€” Congratulations to all co-authors!
  • Dec 15, 2024 β€” Successfully organized the workshop on Handling Resource Constraints using Big Data and AI.
  • Dec 10, 2024 β€” Presented ongoing work at the LC4 Workshop on Real World Applications as part of SPELLL.
  • Jan 10, 2023 β€” Honored to receive the ACL Best Reviewer Prize 2023 β€” Thank you ACL!
  • Sep 10, 2023 β€” Delivered a guest lecture on Multilingual NLP at VIT Chennai.
  • 2022 – Present β€” Mentoring Ph.D. scholars at NIT-K (Hariharan) and VIT Vellore (Ramesh Kannan) on multilingual NLP and Fake News Identification.

🀝 Collaborators

  • Dr. Martin Klignkit (Kyocera Innovation Labs, Japan)
  • Dr. Shinichi Satoh and team (National Institute of Informatics, Japan)
  • Dr. Andreas Dengel, Sheraz Ahmed, Jorn Hees (DFKI Germany)
  • Dr. Snehanshu Saha (APP CAIR, BITS Pilani)
  • Dr. Vinay Namboodiri (DelTA Lab, IIT Kanpur)
  • Dr. Bharathi Raja Asoka Chakravarthi (National University of Ireland, Galway)
  • Dr. Gaurav Sharma (IIT-K / INRIA-THOTH / NEC Media Analytics USA)
  • Dr. Anand Kumar Madasamy (NIT Karnataka, Surathkal)
  • Dr. Sangeetha Sivanesan (NIT Tiruchirapalli)
  • Dr. Ratnavel Rajalakshmi (VIT Vellore)

πŸ“š Selected Publications

  • GISA: Gradual Information Selection Attention for MCQ Difficulty Estimation
    Manikandan Ravikiran, Tarun Sharma, Rajat Verma, Rohit Saluja, Arnav Bhavsar
    AIED 2025 (Core A) – Accepted

  • TEEMIL: Towards Educational MCQ Difficulty Estimation in Indic Languages
    Manikandan Ravikiran, Siddharth Vohra, Rajat Verma, Rohit Saluja, Arnav Bhavsar
    COLING 2025

  • AEI-DRL: Adaptive Ensemble Imputation for Trip Data using Deep Reinforcement Learning
    Ankit Sharma, Akhash Vellandurai, Thiruvengadam Samon, Vinoth Kumar, Manikandan Ravikiran
    IEEE BigData 2024

  • DKT: A First Look at Dynamic Kernel Tuning for Pedestrian Attribute Recognition
    Manikandan Ravikiran, Rahul Mishra, Soumen Biswas, Ananth Ganesh (Equal Contribution)*
    IEEE BigData 2024

  • Hi-GOTE: Hierarchical Groupwise Temporal Ensembling for Pedestrian Attribute Recognition
    Manikandan Ravikiran, Soumen Biswas, Ananth Ganesh
    IEEE ICMLA 2023

  • Revisiting Automatic Speech Recognition for Tamil and Hindi Connected Number Recognition
    Rahul Mishra, Senthil Raja Gunaseela Boopathy, Manikandan Ravikiran, et al.
    Third Workshop on Speech and Language Technologies for Dravidian Languages (2023)

  • Findings of the Second Shared Task on Offensive Span Identification in Code-Mixed Tamil-English Comments
    Manikandan Ravikiran, Bharathi Raja Chakravarthi, et al.
    Third Workshop on Speech and Language Technologies for Dravidian Languages (2023)

  • MMOD-MEME: A Dataset for Multimodal Face Emotion Recognition on Code-Mixed Tamil Memes
    Ramesh Kannan, Manikandan Ravikiran, Ratnavel Rajalakshmi
    LC4 Workshop, CCIS Springer Series (2023)

  • Overlapping Word Removal is All You Need: Revisiting Data Imbalance in Hope Speech Detection
    Hariharan RamakrishnaIyer LekshmiAmmal, Manikandan Ravikiran, et al.
    Journal of Experimental and Theoretical Artificial Intelligence, Taylor and Francis (2023)

  • You Reap What You Sow: Revisiting Intra-Class Variations and Seed Selection in Temporal Ensembling
    Manikandan Ravikiran, Siddarth Vohra, Yuichi Nonaka, et al.
    COMSYS 2021

  • DOSA: Dravidian Code-Mixed Offensive Span Identification Dataset
    Manikandan Ravikiran, Subbiah Annamalai
    Workshop on Dravidian Languages, EACL 2021

  • A Sensitivity Analysis (and Practitioners’ Guide to) of DeepSORT for Low Frame Rate Video
    M. Ravikiran, Y. Nonaka, N. Mariyasagayam
    IEEE Big Data 2020

  • Multilayer Pruning Framework for Compressing Single Shot Multibox Detector
    Pravendra Singh, Manikandan Ravikiran, Neeraj Matyali, Vinay P Namboodiri
    IEEE WACV 2019

(See full list on Google Scholar)

πŸ§ͺ Active Projects

  • Latent Plan Execution for Text Completeness
    Graph-based decoding framework to assess intent alignment in LLM outputs.

  • CASSA / GISA Attention
    Custom attention models for modeling MCQ difficulty using inductive reasoning/psychometric modeling and Analytical Geometry.

  • PEARL for Explanation Auditing
    Post-hoc reasoning traceability for explanation sufficiency evaluation in LLMs.

  • Dialectual Educational Platform
    Developing culturally and linguistically adaptive learning systems for underrepresented languages and dialects.


πŸŽ“ Teaching and Professional Service

Professional Services

  • TPC Member: ACM ICMR (2019), IEEE ICDDS (2022), FIRE (2021, 2022), DravidianLangTech (2024,2023, 2022, 2021), LTEDI (2024,2023, 2022, 2021)
  • Reviewer: ACL (2025, 2024,2023, 2020, 2017), EMNLP (2024, 2023, 2022), AACL (2024, 2022, 2020), EACL (2021, 2023) , AIED(2025), NAACL (2024, 2018), COLING (2024,2018), IEEE Bigdata (2020), ACM ICMR (2019), IEEE ICMLA (2023), IEEE ICDDS (2022), FIRE (2021, 2022), DravidianLangTech (2023, 2022, 2021), LTEDI (2023, 2022, 2021), Springer Language Resources and Evaluation Journal, Elseivier Engineering Applications of Artificial Intelligence, ACM Transactions on Asian and Low-Resource Language Information Processing, Springer Nature Computer Science Journal, Taylor & Francis Journal of Experimental and Theoretical AI
  • Organizer: Workshop on Handling Resource Constraints for/using ML (IEEE Bigdata 2024), Fourth Workshop on Speech and Language Technologies for Dravidian Languages (EACL 2024), Special Session on Machine Learning for Graphs (IEEE ICMLA 2023) , Special Session on Handling Resource Constraints for/using ML (IEEE ICMLA 2023), Workshop on Low Resource Cross-Domain, Cross-Lingual & Cross-Modal Offensive Content Analysis (SPELLL 2022, 2023), Workshop of Cross Modal Learning and Application (ACM ICMR 2019)
  • Mentor: ACM CSCW 2019 Student Reviewer Mentor
  • Chair Roles: Publicity Chair (SPELLL 2022), Session Chair (DravidianLangTech), Industry Session Chair (ICMR)

Patents

  • System and Method Inference Scaling with Weight Reusability
    Manikandan Ravikiran, Ananth Ganesh β€” 2024 (Under Review)

  • System and Method for Finding Novel Objects Across Domains via Weight Switching and Tracing
    Manikandan Ravikiran, Ananth Ganesh, Yuichi Nonaka β€” 2022 (Granted)

  • System and Method for Generalization through Reimann Conditioned Representation
    Manikandan Ravikiran, Yuichi Nonaka, Kingshuk Banerjee β€” 2021 (Under Review)

  • System and Method to Generate Gating Sequences for Training Model on Multidomain Datasets
    Shibashish Sen, Manikandan Ravikiran, Yuichi Nonaka, Nestor Mariyasagayam β€” 2021 (Granted)

  • System and Method for Generating Filter Sequences to Train Model on Completely Noisy Dataset
    Manikandan Ravikiran, Yuichi Nonaka, Nestor Mariyasagayam β€” 2021 (Granted)

  • System and Method to Train Neural Network with Heterogenous Distillation
    Manikandan Ravikiran, Shibashish Sen β€” 2020 (Granted)

  • System and Method to Train Object Recognition Network with Less Data
    Manikandan Ravikiran β€” 2019 (Granted)

  • Method and System for Generating an Optimal Object Recognition Network (OORN) with Balanced Accuracy
    Manikandan Ravikiran β€” 2017 (Granted)

  • Method and System for Determining Plausibility of a Clinical Care Plan
    Manikandan Ravikiran, Sarath P. R., Saima Mohan β€” 2015 (Granted)