AI systems
End-to-end platforms, local LLMs, agentic workflows, reflective safety, Python services, APIs, automation and dependable system architecture.
PythonFlaskAgentsAutomation
Human-centred AI · Systems · Design · Research
AI should extend human creativity — not replace it.
I design and evaluate AI systems that solve real problems while keeping people, judgement and creative control at the centre.
AI systems that keep people in control
Current focus
Agentic systems · LLM evaluation · human-centred automation
Recent work
EDUNEXT, reflective desktop agents, shop-floor intelligence
Background
PhD · industry · research · design
Human-centred AIAgentic systemsInteraction designBehavioural researchIndustrial systemsLearning technologyHuman-centred AIAgentic systemsInteraction designBehavioural researchIndustrial systemsLearning technology
PhDAdvanced Systems Engineering
328+Academic citations
10+ yearsAcross engineering, design & research
Enterprise + academiaTCS, GKN, IISc, NID, PES, unibz
Direction
Build useful AI without reducing human creativity.
My work combines system development, behavioural research and design to make AI more capable, understandable and responsible.
Human behaviour & HCI
Human decision-making, behavioural experiments, usability, cognitive heuristics, recommendation systems and evidence-led design.
HCIPsychologyStatisticsUX research
Industrial systems
ERP/MES data, shop-floor prediction, simulation, human performance, reward mechanisms and Industry 5.0 decision support.
ERP/MESMLSimulationIndustry 5.0
Design & interaction
Physical computing, inclusive design, AR/VR, connected products, prototyping and translating complex technology into usable experiences.
Design thinkingAR/VRIoTPrototyping
Selected work
Projects
Selected systems, experiments and design work across AI, HCI and industrial applications.
EDUNEXT learning intelligence platform
System-level ownership of an AI pipeline that turns lecture video into structured, searchable and personalised learning material.
Reflective offline desktop agent
A privacy-preserving local agent that observes, plans, executes approved actions and independently checks its own work.
Activity-based shop-floor intelligence
Prediction, simulation and human-performance models for flexible manufacturing using real ERP/MES data.
Eco-friendly recommendations & decision heuristics
An experiment on how budget constraints, local cues and price heuristics affect adoption of sustainable recommendations.
Indirect prompting vulnerabilities & conversational defence
A multi-model study of prompts that conceal user intent, paired with an intention-aware defence architecture.
Breathing audio + chest X-ray diagnosis
A deep-learning system combining respiratory sound and radiography to reduce false negatives in COVID-19 screening.
Assistive and accessible technology
Human-centred devices and applications for visual impairment, dementia support, emergency response and rural healthcare.
Demystifying Bharatanatyam
A multimodal framework using labelled gestures, expressions and VR to improve comprehension of Indian classical dance.
Design archive
Earlier product and interaction design work
Products, interfaces and physical-computing prototypes.
Profiles
Research, design and technical work
External profiles with publications, design work, code and academic activity.
Google ScholarPublications, citations and research trajectory.
ORCIDPersistent research identity and publication record.
BehanceDesign archive with products, interfaces and visual work.
GitHubCode, prototypes and technical experiments.
ResearchGateResearch dissemination and scholarly networking.
ACM profileConference and HCI-related academic outputs.
Research
Selected publications
Work across human-centred AI, HCI, industrial systems and multimodal technology.
2025
Reward Mechanisms and Human Performance Management in Activity-Based Shop Floor SystemsPhD dissertation · Free University of Bozen-Bolzano
↗
2025Impact of social rewarding in dynamic shop floor environmentsScientific Reports
↗
2025Prediction of task occurrence distribution for automated shop floor planningProcedia CIRP
↗
2025Demystifying Bharatanatyam: A Multi-Modal FrameworkCHItaly · ACM
↗
2025No- and Low-Alcohol Wines: Perception and AcceptanceFoods
↗
2024Simulation-Driven Evaluation of Reward SystemsDSMIE · Springer
↗
2021A deep-learning multimodal system for COVID-19 diagnosisApplied Soft Computing
↗
2020Design and development of an assistive device for the visually impairedProcedia Computer Science
↗
Experience
Professional background
2026—Now
Independent researcher & AI systems developer
Bolzano, Italy
Reflective desktop agents, LLM evaluation, conversational safety and multidisciplinary research development.2025—2026
Research Fellow / AI Systems Engineer
Free University of Bozen-Bolzano
EDUNEXT system architecture, AI learning pipelines, automation, agentic workflows and cross-team technical coordination.2021—2025
PhD Researcher, Advanced Systems Engineering
Free University of Bozen-Bolzano · GKN Sinter Metals
Human performance, task prediction, ERP/MES analytics, reward mechanisms, simulation and activity-based shop floors.2021—2022
Associate Faculty, Interaction Design
National Institute of Design, Bangalore
Physical computing, multimodal interaction and hybrid environments.2018—2021
Assistant Professor & Lab Head
PES University, Bangalore
Founded and managed the Mechatronics & Interactive Smart Devices Lab; taught HCI, AI, AR/VR, game design and prototyping.2018
Research Assistant
Indian Institute of Science, Bangalore
Urban sustainability, systems analysis and carbon-footprint research.2015—2016
Assistant System Engineer
Tata Consultancy Services
Enterprise manufacturing software, structured delivery, issue resolution and cross-functional teamwork.PhD · Advanced Systems Engineering
Free University of Bozen-Bolzano, Italy
Master of Design · Product Design & Engineering
Indian Institute of Science, Bangalore · Government-funded scholarship
B.Tech · Mechanical & Production Engineering
University of Kerala · Fourth university rank
Approach
Use AI where it adds value. Keep people responsible for meaning, judgement and creativity.
01Frame
Map stakeholders, constraints, data and actual behaviour before choosing a technology.
02Prototype
Connect software, models and interfaces into a testable end-to-end workflow.
03Evaluate
Use experiments, metrics and user evidence rather than relying on technical novelty alone.
04Improve
Add reflection, monitoring and recovery so the system becomes more dependable over time.
Contact
Let’s discuss a role, system or collaboration.
Currently based in Bolzano and open to multidisciplinary opportunities, research collaborations and technology projects in Italy and Europe.