05 / Research · Nitin Mane

Curiosity, engineered.
Purpose, applied.

Exploring the space between software, intelligent systems and real-world needs. From research publications to laboratory and humanitarian collaborations.

Read the publications Explore collaborations
Software meets laboratory automation — event photo 1
A moment from the workSoftware × laboratory automation
↗
2 publicationsProgramming education & vision
Applied researchEdge AI · Robotics · Automation
Cross-disciplineEngineering with real-world context
01 — Peer-Reviewed Publications

2 Papers. Real-World Systems.

Paper 01 - 2025 ACM CODS 2025

Autograder+: A Multi-Faceted AI Framework for Rich Pedagogical Feedback in Programming Education

Vikrant Sahu, Gagan Raj Gupta, Raghav Borikar & Nitin Gautam Mane

Abstract

Automated assessment in programming education typically reduces feedback to a single pass/fail signal, leaving students with little insight into code quality or reasoning. Autograder+ introduces a multi-faceted AI framework that pairs static code analysis — linting, complexity and test-coverage metrics — with semantic feedback from large language models, producing structured, rubric-aligned comments across correctness, style and design. The framework is built to scale formative feedback in large programming courses without sacrificing pedagogical depth.

Cite this work — Vancouver style

Sahu, Vikrant, Gupta, Gagan Raj, Borikar, Raghav, and Mane, Nitin Gautam. Autograder+: A Multi-Faceted AI Framework for Rich Pedagogical Feedback in Programming Education. In: Proceedings of the 13th ACM IKDD International Conference on Data Science (CODS 2025), pp. 306-315. https://doi.org/10.1145/3799830.3799878

View on ACM DL DOI: 10.1145/3799830.3799878
Paper 02 - 2022 SSRN Electronic Journal · 2022

An Improved Object Tracking and Estimation Using Adaptive Kalman-Filter-Based YOLOv5 Model and Mixed Precision for Efficient Inference Performance

Abhilasha Mishra, Nitin Mane, Aaditya Mishra, Amitabh Dixit, Sparsh Mishra, Aditi Agrawal & Rajendraprasad Pagare

98% Training Accuracy
95.4% Test Accuracy
45fps 1080p Inference
14.8 Throughput (mixed precision)
Abstract

Computer-vision and object-detection research is focused on improving human vision and the identification of objects in working scenarios. This research proposes an improved technique using the YOLOv5 model with an adaptive Kalman filter for real-time object tracking and estimation, addressing challenges such as camera alignment, image quality, darkness, occlusion and other environmental factors. The proposed technique achieves better performance by utilising mixed precision for processor selection and parallel distribution of features, demonstrating high model accuracy and real-time processing capability — a training accuracy of 98% and a testing accuracy of 95.4% for surveillance applications, with a throughput of 14.8 and an inference rate of 45 FPS at 1080p resolution. The work contributes to overcoming common tracking challenges and provides solutions for conditional processing cases.

Cite this work — Vancouver style

Mishra, Abhilasha and Mane, Nitin and Mishra, Aaditya and Dixit, Amitabh and Mishra, Sparsh and Agrawal, Aditi and Pagare, Rajendraprasad, An Improved Object Tracking and Estimating Using Adaptive Kalman Filter Based Yolov5 Model and Mixed Precision for Efficient Inference Performance. Available at SSRN: https://ssrn.com/abstract=4187576 or https://dx.doi.org/10.2139/ssrn.4187576

View on SSRN DOI: 10.2139/ssrn.4187576
02 — Current Research Focus

What I'm building at IIT Bhilai.

My M.Tech research extends the tracking-and-estimation line of work above into production-shaped, edge-deployable systems.

01

Industrial Anomaly Detection at the Edge

Extending the Kalman-filter + YOLO line of work into industrial safety monitoring — YOLOv8 pipelines deployed via OpenVINO for real-time Bar & Rod Mill video analytics.

Read Article
02

Multi-Modal LLM/RAG for Plant Intelligence

Fusing vision-model outputs, sensor telemetry and log streams through a retrieval-augmented LLM pipeline that produces plain-language operator alerts in real time.

Read Article
03 — Thesis Contributions

Current M.Tech Research Contributions.

  1. Signal consolidation using multi-modal alert summarization
  2. Intelligent thermal image processing for hot rolling process monitoring
  3. Stable-diffusion-based high-speed camera denoising
  4. Stand-16 looper anomaly detection using Edge AI video analytics
04 — Research Interests

Topics I read, write and experiment with.

Computer Vision Object Detection & Tracking Edge AI / OpenVINO Explainable AI (XAI) Multi-modal LLMs & RAG Industrial IoT Mechatronics & Robotics Healthcare AI Adaptive Filtering Mixed-Precision Inference
Selected experiences

Research beyond the paper.

Collaborations connecting software and engineering to practical needs.

03 STORIES / RESEARCH

3 stories

Open a photo. Explore the story.
Software meets laboratory automation — event photo 1 ↗ View gallery · 05
Software engineering collaborationResearch collaboration

Software meets laboratory automation

Infinity Tech Resources × Institute of Science · Aurangabad

A collaboration exploring software support and robotic automation for chromatography with biotechnology researchers.

Explore the story

The work

At Infinity Tech Resources, participated in a research environment connecting software engineering with the Government Institute of Science Aurangabad’s Department of Biotechnology.

The contribution

Brought software and automation perspectives into discussions with biotechnology researchers about chromatography systems.

Link to this story
Engineering for a local healthcare need — event photo 1 ↗ View gallery · 04
Project contributor2021

Engineering for a local healthcare need

IEEE HAC / SIGHT · MIT Hospital, Aurangabad

Contributed to a Smart Sterilizer Box for masks and PPE kits under IEEE’s “Addressing Local Needs During COVID Time” initiative.

Explore the story

The work

Worked with Dr. Abhilasha Dixit Mishra and the project team, supported by USD 450 in IEEE HAC/SIGHT and Region 10 funding.

The contribution

The project progressed from prototyping to a unit delivered to MIT Hospital, with input from medical staff. The completion certificate recognizes the team’s contribution.

Link to this story
Nitin Mane in a safety helmet beside the rolling line at the Bar & Rod Mill, Bhilai Steel Plant ↗ View gallery · 04
M.Tech researcherMar & Oct 2025

Field research at the Bar & Rod Mill

Bhilai Steel Plant (SAIL)

Bhilai, Chhattisgarh

RelatedRead the case studyIIT Bhilai

Site visits to the Bar & Rod Mill at SAIL's Bhilai Steel Plant, where the industrial anomaly-detection research for my M.Tech at IIT Bhilai is grounded.

Explore the story

The work

Seeing the rolling line and hot bars on the roller tables in person informs how the detection pipeline must cope with heat glare, dust, motion and occlusion on a real plant floor.

The contribution

Field observations feed directly into the edge-AI anomaly-detection case study.

Link to this story
Collaborate

Working on something
similar?

Open to co-authoring, peer review, or just trading notes on real-time tracking, edge deployment or multi-modal systems.

Open image ↗