Research

I have a broad interest in training and inference optimization techniques, with a particular focus on applications in the vision and language domains. I am passionate about exploring advanced methods to improve model efficiency and performance in these cutting-edge fields.

Google Scholar

📚 Publications

🗣️ Language

Unlocking the Edge Deployment and On-Device Acceleration of Multi-LoRA Enabled One-for-All Foundational LLM

Sravanth Kodavanti, Sowmya Vajrala, Srinivas Miriyala, Utsav Tiwari, Uttam Kumar, Utkarsh Kumar Mahawar, Achal Pratap Singh, Arya D, Narendra Mutyala, Vikram Nelvoy Rajendiran, Sharan Kumar Allur, Euntaik Lee, Dohyoung Kim, HyeonSu Lee, Gyusung Cho, JungBae Kim

ACL 2026 Findings

Paper Poster Slides Video

👁️ Vision

TOC-SR: Task-Optimal Compact diffusion for Image Super Resolution

Sowmya Vajrala, Akshay Bankar, Manjunath Arveti, Shreyas Pandith, Sravanth Kodavanti, Subhajit Sanyal, Amit Unde, Srinivas Miriyala

LoViF, ECCV 2026 (Oral)

Paper

NanoSD: Edge Efficient Foundation Model for Real Time Image Restoration

Subhajit Sanyal, Srinivas Miriyala, Akshay Bankar, Manjunath Arveti, Sowmya Vajrala, Shreyas Pandith, Sravanth Kodavanti, Abhishek Ameta, Harshit, Amit Unde

CVPR 2026

Paper Poster Video

EdgeDiT: Hardware-Aware Diffusion Transformers for Efficient On-Device Image Generation

Sravanth Kodavanti, Manjunath Arveti, Sowmya Vajrala, Srinivas Miriyala, Vikram N R

Mobile AI, CVPR 2026 (Oral)

Paper

Hardware-aware Low Light Image Enhancement on Edge

Sowmya Vajrala, Sravanth Kodavanti, Srinivas Miriyala

WiCV, CVPR 2026 (Spotlight)

Paper

Quantization with Unified Adaptive Distillation to enable multi-LoRA based one-for-all Generative Vision Models on edge

Sowmya Vajrala, Aakash Parmar, Prasanna R, Sravanth Kodavanti, Manjunath Arveti, Srinivas Soumitri Miriyala, Ashok Senapati

Mobile AI, CVPR 2026 (Oral)

Paper

Mobile-friendly Image de-noising: Hardware Conscious Optimization for Edge Application

Srinivas Miriyala, Sowmya Vajrala, Hitesh Kumar, Sravanth Kodavanti, Vikram N R

ICASSP 2025

Paper Poster

Edge-Efficient Image Restoration: Transformer Distillation into State-Space Models

Sravanth Kodavanti, Srinivas Miriyala, Sowmya Vajrala, Vikram N R, Sharan Allur

Preprint

Paper

Towards Efficient Image Deblurring for Edge Deployment

Sowmya Vajrala, Srinivas Miriyala, Sravanth Kodavanti

Preprint

Paper

⚙️ Patents

Multi-objective Optimization for Balanced Inference-time Latency-efficient and Eviction-aware SWA (Mobile SWA)

Sravanth Kodavanti, Srinivas Miriyala, Sowmya Vajrala, Nikita Malik, Vikram N R

Provisional Specification Filed

Proposes a dynamic window for generic token eviction algorithms reducing KV Cache without significant drop in accuracy of LLMs.

System and Method for Accelerating On-Device Large Language Model Inference

Sravanth Kodavanti, Srinivas Miriyala, Sowmya Vajrala

Complete Specification Filed

Proposes a novel self-speculative decoding technique to significantly improve efficiency and reduce latency in LLM inference.

🎓 Thesis

Continual Learning for 3D Point Cloud

Under guidance of Dr. P K Srijith

IIT Hyderabad (2023–2024)

Poster