AI/ML Researcher & Systems Architect
Engineering practical, explainable intelligence and resilient infrastructure — bridging the frontier between deep learning models, vision-language systems, and production full-stack platforms.
I'm a B.Tech Computer Science (Artificial Intelligence) student at Adi Shankara Institute of Engineering and Technology, Kerala (CGPA: 8.81), focused on building practical, user-centered AI systems — spanning explainable deep learning, computer vision, and resilient full-stack platforms.
I recently completed an intensive 3-month NPTEL Summer Internship at IIT Kharagpur under Prof. Priyadarshi Patnaik (with technical guidance from Ankit Kumar and Madhur S. Pradhan), investigating tradition-consistency in Indian craft objects through machine learning and vision-language models. This delivered three major milestones: Anom-Reason (a PatchCore + Grad-CAM + Gemini reasoning proof-of-concept achieving AUROC up to 0.923 on a Ghibli substitute domain), the Tradition Annotation Pipeline (a human-in-the-loop dataset curation system structuring 369 images across 11 traditional Indian art styles), and Trad-Reason (a deployed conversational assessment agent providing grounded consistency verdicts over the approved corpus).
On the research front, I co-authored a peer-reviewed research paper presented and published with IEEE at ICIMRBE 2026, introducing a bidirectional cross-attention fusion architecture with SHAP Deep Explainer interpretability for multimodal vital signal anomaly detection (achieving 99.7% ensemble accuracy). Currently, my college major project focuses on Adaptive Vision Defect Detection — uniting hierarchical CNN and Vision Transformer (Swin) backbones with Gemini VLM diagnostic reasoning for explainable surface defect localization.
My practical foundation includes a Winter Internship at IIT Ropar (Vicharanashala Lab under Prof. Sudarshan Iyengar) optimizing real-time AI question pipelines, qualifying as a National Finalist at EnCode 2026 (IIT Guwahati, Top 10 in India), and achieving LeetCode Knight status (peak contest rating 2,077, placing in the Top 1.75% globally). In my 3rd year, I also qualified GATE 2026 (CS) with a 399 score (Top 8.8% nationally) purely through hands-on practice and self-study.
As an active open-source creator, I built Bastion, a privacy-focused GNOME Shell extension with 400+ downloads on GNOME Extensions. From architecting national-level CTF competitions and technical debugging events (Prayag '25 & '26) to maintaining a virtualized home lab running self-hosted Linux containers, Cloudflare Tunnels, and automated DevOps workflows, I am dedicated to engineering impactful, production-grade technology.
NPTEL Summer Internship 2026 (Completed) - Completed Summer Internship at IIT Kharagpur under the guidance of Prof. Priyadarshi Patnaik, working alongside researchers Ankit Kumar (technical guidance) and Madhur S. Pradhan (science behind craft). Developed machine learning and vision-language models to assess whether an image, artwork, or craft object aligns with an Indian visual tradition across three completed phases: developing the Anom-Reason proof-of-concept (ResNet-34 + PatchCore + Gemini reasoning), engineering the human-in-the-loop Tradition Annotation Pipeline for a 369-image eleven-style Indian art dataset, and deploying Trad-Reason, a grounded conversational assessment agent with live consistency verdicts.
Duration: July 12, 2026 - September 2026 (Completed) | Focus: Tradition-Sensitive Visual Assessment using ML/Vision-Language Models | Mode: Virtual
View Project ProgressNPTEL Winter Internship 2025 - Completed Winter Internship at IIT Ropar Under the guidance of Prof. Sudarshan Iyengar, focusing on full-stack development using the MERN stack to build scalable, high-performance web architectures that complement complex backend systems. Selected for this prestigious 4-week virtual internship program based on Elite + Silver performance (Top 5%) in "Joy of Computing Using Python" course.
Duration: December 22, 2025 - January 18, 2026 | Focus: MERN Stack Development | Mode: Virtual
View Weekly ProgressSupervised & Unsupervised Learning, Reinforcement Learning, ML Algorithms
AI Fundamentals, Ethical AI, Computer Vision
Data Analysis, Data Visualization, Data Science
IBM Watson, Watson Studio, OpenCV
Problem Solving, Scripting, ML/AI Development
OOP, JDBC, Java Swing, Desktop Applications
System Programming, Game Development, DSA
Web Development, Frontend & Backend
Web Design, Responsive Layouts
Database Queries, Data Management
React.js, JavaScript, Responsive Design
Node.js, API Development, REST APIs
End-to-end Web Applications, Hosting & Deployment
Tailwind CSS, Data Pipelines
Server Administration, Bash Scripting
System Maintenance, Infrastructure Management
Docker, Container Orchestration
Ansible, Cloud Deployment, CI/CD
Prometheus, Grafana, System Monitoring
Git, GitHub, Collaborative Development
Authors: Amrutha Muralidharan Nair, Meenatchi K V, Asha Rose Thomas, Ashna Thajudheen, Devika P Sajith, Gokul P
2026 2nd International Conference on Innovations in Mechanical, Robotics, Computing and Biomedical Engineering (ICIMRBE) · Published by IEEE · Ernakulam, India
Designed for Internet of Medical Things (IoMT) environments, this framework moves beyond single-sensor telemetry to capture clinically rich physiological context by fusing two distinct modalities: real-time vital sign readings (Heart Rate, SpO2, Temperature) and 9-axis time series from Human Activity Recognition (HAR) sensors. A patient's heart rate carries entirely different clinical meaning at rest versus during physical exertion — this framework directly captures and models that distinction.
B.Tech in Computer Science and Engineering (Artificial Intelligence)
Adi Shankara Institute of Engineering and Technology, Kerala
Expected Graduation: May 2027
9.59
Semester 4
8.81
Up to Semester 6
Beyond building software and conducting research, I actively contribute to the technical community through invited talks, IEEE sessions, cybersecurity competitions, and educational initiatives. I enjoy making complex engineering concepts accessible, mentoring fellow students, and creating hands-on learning experiences that encourage practical problem solving.
IEEE Computer Society Kerala Chapter · IEEE Young Professionals Kerala Section · IEEE CS SBC Toc-H
14 July 2026 · Invited Speaker
IEEE technical session exploring how ML models should be evaluated, interpreted, and trusted beyond simple accuracy metrics. Covered regression and classification metrics, ROC-AUC vs PR-AUC, probability calibration, explainability with SHAP, and metric selection for real-world AI systems. Included an interactive walkthrough of a Network Intrusion Detection pipeline trained on the CTU-13 dataset.
GitHub ResourcesIEEE CS Sahrdaya SB · in collaboration with IEEE Computer Society Kerala Chapter & IEEE Young Professionals Kerala Section
28 July 2026 · Invited Speaker
IEEE session focused on one of the most overlooked aspects of machine learning — datasets. Covered discovering, evaluating, collecting, cleaning, annotating, and validating datasets while discussing bias, data quality, preprocessing pipelines, and practical strategies for building reliable AI systems from the data upwards.
GitHub ResourcesDepartment of Artificial Intelligence & Data Science · Adi Shankara Institute of Science and Technology
September 2025 · CTF Architect & Infrastructure Lead
Architected and executed the flagship CTF competition during Prayag '25. Designed the complete challenge set, built the self-contained competition infrastructure, and deployed and hosted the event live. Collaborated with organizers, faculty coordinators, and the technical support team to deliver a hands-on 90-minute cybersecurity competition.
LinkedIn PostAdi Shankara Institute of Science and Technology
February 2026 · CTF Organizer & CTF Architect
Organized and hosted the CTF competition during Aswamedha 2026. Designed cybersecurity challenges, managed the competition infrastructure, and supported participants throughout the event — promoting practical cybersecurity education through hands-on problem solving.
Department of Artificial Intelligence & Data Science · Adi Shankara Institute of Engineering and Technology
Event Organizer & Technical Coordinator
Organized and coordinated Code Conquest, the signature competitive programming and debugging challenge at Prayag '26, ASIET's national-level technical fest. Designed problem statements across algorithmic logic, edge-case analysis, and bug hunting; managed the competition flow and live evaluation environment testing problem-solving agility under pressure.
LinkedIn PostGraduate Aptitude Test in Engineering
Issued Feb 2026 · Score: 399 · Top 8.8%
Skills: Algorithms, Data Structures, Operating Systems, Computer Networks, Theory of Computation, Database Management Systems
NPTEL - Elite + Gold (Top 5%)
IIT Ropar | Issued MAY 2026 · Credential ID: NPTEL26CS66S262700348
Skills: Deep Learning, Neural Networks, Convolutional Neural Networks, Machine Learning
Stanford University & DeepLearning.AI
Issued Dec 2025
Skills: Supervised Learning, Unsupervised Learning, Machine Learning Algorithms, Reinforcement Learning
NPTEL - Elite + Silver (Top 5%)
Issued Nov 2025 • Credential ID: NPTEL25GE55S1265800468
Skills: Machine Learning, Artificial Intelligence (AI)
NPTEL - Elite + Silver (Top 5%)
Issued Apr 2025 • Credential ID: NPTEL25CS69S1054100101
Skills: Problem Solving, Python (Programming Language)
Team Kairos | Top 10 in India
UDGAM, Indian Institute of Technology, Guwahati
Qualified as one of the top 10 teams nationwide for the flagship AI hackathon at IIT Guwahati. Team: Gokul P, Keerthana Vinod, C Raniya Nazrine
What started as a simple idea, a few uncertain conversations, and a shared dream of building something meaningful has now led us here. Competing among thousands of teams across the country and making it to the National Finals is something we'll always be proud of.
From moments of self-doubt to moments of clarity… From late-night debugging to early-morning hope… From "will this even work?" to "we're NATIONAL finalists." This journey gave us more than a project. It gave us growth, resilience, friendship, and belief.
Proud to announce that Team Kairos — Gokul P, Keerthana Vinod, and C Raniya Nazrine — are now National Finalists at EnCode 2026: Code To Innovate, the flagship AI hackathon by UDGAM, IIT Guwahati.
We'd like to sincerely thank everyone who supported us along the way—our parents, Adi Shankara Institute of Engineering and Technology, mentors, friends, and everyone who believed in us when we doubted ourselves. Your constant encouragement, patience, sacrifices, feedback, and motivation played a huge role in bringing us here. This wouldn't have been possible without you.
Now, we move on to the offline finals at IIT Guwahati, ready to push harder, think bigger, and give it everything we've got. No matter what happens next, this journey itself is a victory for us. Onwards and upwards. Let's make this count, Team Kairos. 💙🔥
Score: 399 · Top 8.8% Nationally
February 2026 · Adi Shankara Institute of Engineering and Technology, Kalady
Qualified during 3rd year of B.Tech through self-study and practical engineering experience, without formal coaching.
Successfully qualified GATE 2026 in Computer Science and Information Technology (CS) with a score of 399, placing in the Top 8.8% of all candidates nationally. This milestone was achieved during my 3rd year of B.Tech at Adi Shankara Institute of Engineering and Technology, Kalady.
Rather than following conventional coaching routes, I relied entirely on practical development experience and a strong foundation in core engineering concepts built through hands-on projects, NPTEL courses, and competitive programming.
Explainable AI & Multimodal Physiological Anomaly Detection
ICIMRBE 2026 · IEEE Xplore
Cross-attention-based multimodal physiological anomaly detection with explainable AI; published by IEEE and presented at the International Conference on Innovations in Mechanical, Robotics, Computing and Biomedical Engineering.
Cross-Attention Fusion Based Multimodal Vital Signal Anomaly Detection Using Deep Learning
Amrutha Muralidharan Nair, Meenatchi K V, Asha Rose Thomas, Ashna Thajudheen, Devika P Sajith, Gokul P
The work combines multimodal physiological signal analysis (vital signs + 9-axis HAR IMU telemetry) with bidirectional cross-attention mechanisms to detect anomalies, paired with SHAP Deep Explainer explainable AI so that flagged anomalies come with an interpretable clinical rationale rather than an opaque prediction. Achieved 95.4% standalone accuracy and 99.7% with a 5-model soft ensemble.
Publishing original research with IEEE at an international conference validates the explainable-AI and anomaly-detection methodology running through other research and internship work, applied here to the Internet of Medical Things (IoMT) domain.
Peak Contest Rating 2,077
Top 1.75% globally
Achieved LeetCode Knight status with a peak contest rating of 2,077, placing in the top 1.75% of competitive programmers globally.
Reflects sustained practice in data structures, algorithms, and problem-solving under contest time pressure — the same competitive-programming foundation drawn on for IEEE Xtreme and other timed technical challenges.
Consistent contest performance over time, rather than a single event, demonstrates depth in algorithmic thinking that complements the applied AI/ML and systems work shown elsewhere on the site.
Transitioned to Project-Based Work
IIT Ropar | Vicharanashala Lab for Education Design
Recognized among 6 interns (from several hundred) who successfully completed learning and evaluation phases to transition to real-world project work.
Featured in Vicharanashala Lab's official LinkedIn progress update as one of the select interns who successfully completed the rigorous learning and evaluation phases of the NPTEL-aligned internship program. Out of several hundred virtual interns, only 6 were highlighted for transitioning to project-based work.
This recognition came from Prof. Sudarshan Iyengar and the Vicharanashala Lab team at IIT Ropar, acknowledging the successful completion of intensive learning modules and evaluation criteria. Being among the featured interns demonstrates consistent performance and readiness for real-world application of skills.
Transitioned to applying learning to real-world projects under the guidance of Prof. Sudarshan Iyengar, Pushpendra P. Singh, and the Vicharanashala Lab team, focusing on AI-driven systems and educational technology solutions.
IIT Ropar | Vicharanashala Lab for Education Design
Under Prof. Sudarshan Iyengar
Selected for prestigious 4-week virtual internship based on Elite + Silver performance (Top 5%) in "Joy of Computing Using Python" - IIT Ropar.
Selected for this highly competitive internship program at IIT Ropar's Vicharanashala Lab for Education Design based on exceptional performance in NPTEL courses. The selection was merit-based, requiring Elite + Silver certification (Top 5% ranking) in "Joy of Computing Using Python" taught by Prof. Sudarshan Iyengar.
Worked on AI-driven systems and full-stack development with the MERN stack under the guidance of Prof. Sudarshan Iyengar. The internship involved optimizing real-time question generation pipelines, contributing to educational technology solutions, and participating in peer evaluation activities.
Part of a rigorous NPTEL-aligned virtual internship program with several hundred participants, focusing on upskilling through peer learning and learning-by-doing methodologies. The program emphasized practical application of theoretical knowledge in real-world project scenarios.
Worked under the guidance of Prof. Sudarshan Iyengar, Pushpendra P. Singh, Harshdeep R., Rajan Gupta, Meenakshi V, Mukund Aditya BV, and Nandan Prabhudesai at the Vicharanashala Lab for Education Design, IIT Ropar.
Stanford University & DeepLearning.AI
December 2025
3-course series covering supervised learning, unsupervised learning, and advanced algorithms. Hands-on experience with Python, TensorFlow, and scikit-learn.
After an immersive learning journey, I successfully completed this world-class program offered by DeepLearning.AI and Stanford University on Coursera. This comprehensive 3-course series demanded a deep dive into the core of AI, covering everything from Supervised and Unsupervised Learning to Advanced Learning Algorithms.
Gained hands-on experience building and training models using Python, TensorFlow, and scikit-learn, moving beyond theory to practical implementation of Neural Networks, Decision Trees, and Recommender Systems.
Receiving this certification from Stanford Online represents a comprehensive understanding of the mathematical underpinnings and best practices used in Silicon Valley for AI innovation. The structured learning significantly solidified my technical foundation, preparing me to tackle more complex challenges in Data Science and AI.
Shortlisted - Team Ayunex
Ministry of Education's Innovation Cell
Recognized for innovative project proposal in national-level innovation challenge.
Ayunex has been honored to be shortlisted for BOTH the AICTE Productization Fellowship (APF) and the YUKTI Innovation Challenge (YIC) 2025. This dual recognition represents a significant milestone for our healthcare innovation.
Our project focuses on elderly healthcare innovation, addressing critical gaps in healthcare technology. The recognition from the Ministry of Education's Innovation Cell validates our approach to making healthcare more accessible and effective.
IIT Ropar
Prof. Sudarshan Iyengar
Achieved Top 5% ranking in computational thinking and creative problem-solving with Python.
Secured a spot in the Top 5% (Elite + Silver) in this course conducted by Prof. Sudarshan Iyengar at IIT Ropar. This course was an incredible journey into computational thinking and creative problem solving with Python, truly living up to its name - "Joy of Computing."
This exceptional performance led to selection for the prestigious NPTEL Winter Internship 2025 at IIT Ropar's Vicharanashala Lab for Education Design, where I worked under Prof. Sudarshan Iyengar on AI-driven systems and educational technology.
The course significantly strengthened computational thinking skills and Python programming capabilities, providing a solid foundation for advanced AI and data science work. The creative problem-solving approach emphasized in the course has been invaluable in tackling real-world challenges.
IIT Ropar
12-Week Deep Learning Course
Achieved Top 5% ranking with Elite + Gold certification in Deep Learning from IIT Ropar.
Secured Elite + Gold certification (Top 5%) in the Deep Learning course offered by IIT Ropar through NPTEL. This comprehensive course provided in-depth knowledge of neural network architectures, training methodologies, and real-world deep learning applications.
The curriculum covered foundational and advanced deep learning topics including feedforward networks, CNNs, RNNs, optimization techniques, regularization, and modern architectures. Both theoretical understanding and practical implementation using Python and leading frameworks were emphasized.
This achievement reflects a strong grasp of deep learning principles directly applicable to AI model development, computer vision, and NLP — areas central to my ongoing projects and research interests.
IIT Guwahati
12-Week Rigorous AI Course
Achieved Top 5% ranking among 3,795 certified candidates with 83% consolidated score.
Completed a rigorous 12-week journey in AI fundamentals offered by IIT Guwahati, achieving Elite certification with an 83% consolidated score. The performance secured a spot in the Top 5% of all certified candidates out of 3,795 participants.
The comprehensive curriculum covered core AI concepts including search algorithms, knowledge representation, machine learning fundamentals, neural networks, and intelligent agents. The course provided both theoretical depth and practical implementation experience.
This achievement represents a strong foundation in AI algorithms and methodologies, preparing for advanced work in machine learning, deep learning, and AI-driven systems. The structured learning significantly enhanced understanding of AI principles used in modern applications.