Dr. Rajesh Kumar Sharma
Professor & Head, Department of Computer Science
Indian Institute of Technology, DelhiResearch Areas
Research Overview
My research programme sits at the intersection of theory and deployment — I care as much about proving an algorithm's correctness as I do about seeing it run on real hardware, in real hospitals, farms, and traffic systems. Over 24 years, this has grown from a single thread of doctoral work in distributed consensus into a six-area programme spanning machine learning, computer vision, systems, security, and — most recently — language technology for Indian languages. Below is a detailed look at each area, the work behind it, and where it's headed next.
Machine Learning & Federated Learning
Privacy-preserving distributed learning techniques that allow multiple institutions to train shared models without exposing raw data — with a particular focus on healthcare applications where data sensitivity is paramount.
Key Work & Outcomes
- Principal Investigator on the DST-funded project "Privacy-Preserving Federated Learning for Rural Healthcare" (2023–2026)
- Framework piloted with 6 partner hospitals through a collaboration with AIIMS Delhi for federated diagnostic model training
- 18 peer-reviewed publications in this area, including IEEE TNNLS and NeurIPS
Computer Vision
Real-time object detection and recognition systems engineered to run efficiently on low-power, resource-constrained edge devices — bridging the gap between state-of-the-art accuracy and real-world deployability.
Key Work & Outcomes
- Co-developed the object detection framework behind an AI-driven traffic management system, deployed with the Ministry of Road Transport
- Research on model compression enabled deployment on Raspberry Pi-class hardware without significant accuracy loss
- Work published at CVPR, ECCV, and the International Journal of Computer Vision
Distributed Systems
Fault-tolerant consensus algorithms and scalable architectures for large-scale distributed systems — the area of my doctoral research and a continuous thread throughout my career, from academic theory to industry-adopted protocols.
Key Work & Outcomes
- Byzantine fault-tolerant consensus protocols co-developed with industry partner TechCorp Labs, now used in their production blockchain systems
- Authored "Distributed Systems Design: Principles and Paradigms", adopted as a core text in 40+ Indian university courses
- 20+ years of continuous publication record in this area, from doctoral work through to 2025
IoT Security
Lightweight cryptographic protocols and intrusion detection systems designed for the tight power, memory, and bandwidth constraints of Internet-of-Things devices — with a strong emphasis on real-world field deployment.
Key Work & Outcomes
- Principal Investigator on the MeitY-funded "Secure IoT Framework for Smart Agriculture", field-deployed across 200+ farms in Punjab and Haryana
- Two patents granted for lightweight authentication protocols suited to constrained IoT hardware
- Intrusion detection research for Industrial IoT published in IEEE Internet of Things Journal
Natural Language Processing
Language modeling for low-resource Indian regional languages, addressing the gap between NLP research (largely English-centric) and the linguistic diversity of India's 22 scheduled languages.
Key Work & Outcomes
- Currently supervising a PhD scholar (Neha Kulkarni, expected 2028) on low-resource language modeling for Indian regional languages
- Early collaboration discussions underway with the Bhashini national language initiative
- Newest addition to the lab's research portfolio, building on the group's federated learning expertise for privacy-preserving language data collection
Cloud Computing
Scalable data replication and consistency strategies for applications distributed across geographically dispersed cloud regions, balancing latency, availability, and consistency guarantees.
Key Work & Outcomes
- Research on geo-distributed database replication strategies published in IEEE Transactions on Parallel and Distributed Systems
- Findings informed regional cloud deployment strategy discussions with industry consulting partners
- Natural extension of the group's distributed systems research into modern cloud-native architectures
Funded Research Projects
| Project Title | Funding Agency | Duration | Role |
|---|---|---|---|
| Privacy-Preserving Federated Learning for Rural Healthcare | DST | 2023 – 2026 | Principal Investigator |
| Secure IoT Framework for Smart Agriculture | MeitY | 2022 – 2025 | Principal Investigator |
| Scalable Consensus Protocols for Distributed Ledgers | Industry Partner (TechCorp Labs) | 2021 – 2024 | Co-Principal Investigator |
| Low-Power Computer Vision for Edge Devices | DST | 2020 – 2023 | Principal Investigator |
| AI-Driven Traffic Management System | Ministry of Road Transport | 2018 – 2021 | Co-Investigator |
Distributed Systems & AI Research Lab
I lead the Distributed Systems & AI Research Lab at IIT Delhi, established in 2014. The lab currently hosts 5 doctoral scholars and several master's students working across our core research areas, with active collaborations with industry partners and international universities including the National University of Singapore and ETH Zurich.
International Collaborations
Cross-border partnerships have shaped several of the lab's larger projects, from joint publications to co-supervised research visits.
- National University of Singapore — joint work on Byzantine fault-tolerant consensus, 2021–present
- ETH Zurich — collaborative research on federated learning for healthcare, 2023–present
- University of Sydney — co-authored work on IoT security presented at ACM SIGCOMM, 2023