I build computational models of
how audiences watch films.

I study how audiences perceive, interpret and remember cinematic media, combining computational perception models, narrative structure analysis and interactive media systems to understand how films guide attention, emotion and interpretation.

My work models how viewers engage with moving-image media using behavioural analysis, saliency modelling, temporal media analytics and AI-assisted narrative systems. I treat attention as a perceptual mechanism and trust as a higher-level interpretive judgement, and study how computational media systems shape the relationship between the two.

London, United Kingdom · Open to research collaboration and PhD conversations

02Recent

2026
Two papers under review: When Transparency Hurts and Helps (trust in AI-generated news) and Computational Tholpavakkoothu (human–AI dramaturgy).
2026
Invited lecture at the Creative Computing Institute, UAL: The Mathematics of Cinema: Patterns, Structure, and Perception in Film.
2026
Presented DinoSynthesis at the UAL × CHEAD Generative AI Event, London.
2026
Runner-up, Makerversity Under 30s, Somerset House, London; second place, UAL CES × Lovable Portfolio Showcase Gala.
2025
Began the MRes in Creative Computing at UAL with a competitive £25,000 scholarship.

03Research

The long-term goal of my research is to build computational models of cinematic perception: artificial systems that can watch films, identify narrative structures, and model viewer engagement in ways comparable to human audiences.

Computational film perception

How visual attention, emotion and narrative structure shape viewer engagement, studied through behavioural signals, saliency modelling and temporal media analytics.
Narrative structure intelligence

Computational models that extract structural patterns (phases, transitions, tension, causal density) from film scripts and cinematic sequences, and relate narrative form to audience reception.
Generative and interactive media systems

Experimental systems exploring new forms of storytelling, performance and AI-mediated media, including computational treatments of traditional performance forms and questions of trust in autonomous media.

04Education

2025–present

MRes in Creative Computing, Creative Computing Institute, University of the Arts London

£25,000 competitive scholarship. Research focus: computational media, cinematic perception, narrative intelligence, multimodal AI systems.
2022–2024

MTech in Computer Science & Engineering (Artificial Intelligence), Digital University Kerala

Outstanding Student Award, 2022–2024 batch. Research focus: large language models, computer vision, multimodal AI systems.
2021–2022

PG Diploma in Data Science and Analytics, Kannur University
2019–2021

MSc Physics, Kannur University

05Experience

Dec 2024–Sept 2025

AI Research & Development Engineer, Digital University Kerala

Engineered multimodal AI systems spanning large language models, computer vision pipelines and AI-assisted verification; built LLM pipelines, multi-agent workflows and retrieval-augmented generation architectures; automated data extraction and multimodal information processing for digital media and verification tasks.
Jun–Sept 2025

AI Academic Tutor, ASAP Kerala

Structured teaching in AI, Python, machine learning and computational thinking; workshops and mentoring for students and working professionals.
Jul 2024–Sept 2025

Co-Founder, KnowLumi

AI-driven education platform built around adaptive interview simulation; designed and deployed a multi-agent AI interviewer; led AI workflow architecture, experimentation and deployment.
Jul–Dec 2023

Research Intern, Indian Institute of Management Kozhikode

Large language models and enterprise information systems; prompt engineering workflows and AI-assisted information retrieval.
Apr–Jun 2021

Research Intern, CSIR-National Physical Laboratory, New Delhi

Atmospheric analysis using NASA AERONET datasets; statistical analysis pipelines and scientific reporting.

Teaching

2018–present

Independent Mathematics & Physics Tutor

Mentored more than 2,000 students across physics and mathematics for higher-secondary and university level, including IIT-JEE and NEET preparation.
2022–present

Coding & AI Instructor

Beginner-friendly sessions in Python, AI and computational thinking for students from diverse academic backgrounds.

Creative industry

Collaborated on approximately 15 short films, advertisements and music video productions across assistant direction, creative coordination, editing support and technical production, coordinating multidisciplinary pipelines and translating between creative and technical teams.

06Publications

Peer-reviewed

Rahulraj P. V., Sanil J., Anoop V. S., Asharaf S. (2023). Monetize the Dual: A Data-Analytic Approach for Native Language and Prequel Movie Popularity Analysis. Proceedings of the International Conference on Data Analytics and Insights (ICDAI), Springer. Springer

Predicting film popularity by analysing linguistic origin and franchise lineage using multi-modal data streams.

Under review

Rahulraj P. V., Sanil J., Anoop V. S., Asharaf S. (2026). Narrative DNA: DNA-Sequence Inspired AI-Driven Analysis of Narrative Dynamics in Movie Scripts. Society for Cognitive Studies of the Moving Image (SCSMI). Draft

Computational analysis of film scripts to extract narrative phases, sub-genre transitions and structural patterns, studying how narrative form relates to audience reception.
Rahulraj P. V. (2026). When Transparency Hurts and Helps: Trust in AI-Generated News.

Examines how users interpret, trust and reason with AI-mediated news, using a multi-agent newsroom system (Gradient.ai) as an experimental apparatus that surfaces bias, uncertainty and perspective through multiple algorithmic framings.
Rahulraj P. V. (2026). Computational Tholpavakkoothu: Human–AI Dramaturgy for Living Shadow Rituals.

A computational shadow-puppetry system inspired by Tholpavakkoothu, examining joint-based motion grammars and semantic light control for performative computation.

Earlier work · physics & education

Rahulraj P. V., Antony E. (2022). Changes in Atmospheric Air Quality in the Wake of a Lockdown Related to COVID-19 in Kerala. International Journal of Agriculture & Environmental Science (IJAES).
Rahulraj P. V., Antony E. (2022). Influence of AI in Education System. Emerging Trends of Psycho-Technological Approaches in Heutagogy, TNOU.

07Research systems

Systems built as research instruments. Expand any entry.

Saliency model output over a two-shot from Vidheyan (1994, dir. Adoor Gopalakrishnan). Predicted attention settles on the two faces and shifts between them across the shot. Representative output from the perception pipeline underlying SPECT.

SPECT · Computational Spectatorship Modelling 2025–26

The umbrella framework for my computational spectatorship research: modelling viewer perception, attention and narrative understanding in film as one system rather than as separate measurements. It combines saliency modelling with multimodal perception signals and behavioural experimentation, and gives the individual studies (engagement analysis, attention visualisation, narrative topology) a shared structure to report into.

Computational modelling of audience attention across a runtime
Narrative engagement and interpretation modelling
Saliency systems driven by multimodal perception signals
Behavioural experimentation as empirical grounding

Methods: saliency modelling, multimodal AI, behavioural experimentation, perception analytics. Python.

3D-MSPI Attention Visualisation Pipeline 2025–26

Most saliency work flattens a film to a 2D heatmap, which discards the dimension cinema is composed in: depth. This pipeline augments saliency with depth estimation to extract attention as a trajectory through 3D space, then clusters those trajectories to find where perception groups and where it scatters across a sequence.

Depth-augmented saliency extraction
3D attention trajectory reconstruction across sequences
Perceptual clustering of attention paths

Methods: computer vision, saliency modelling, depth estimation, clustering. Python.

Narrative DNA 2023–26

Borrows its method from sequence analysis in biology. Film scripts are encoded as symbolic sequences, then mined for recurring phases, transitions and structural motifs, the way a genome is read for genes and regulatory patterns. The output is a structural fingerprint of a film that can be compared across a corpus and correlated against audience reception. Basis of the SCSMI 2026 submission.

Symbolic encoding of screenplay structure into analysable sequences
Narrative phase extraction and sub-genre transition detection
Structural motif mining across a film corpus

Methods: sequence analysis, NLP, clustering. Python. Paper draft

Gradient.ai · Autonomous News Intelligence & Trust System 2025–26

A multi-agent news system designed to surface bias, uncertainty and perspective by presenting multiple algorithmic framings of the same story. It is both a working pipeline (aggregation, cross-checking, transparency-oriented validation, algorithmically assisted broadcast generation) and a research instrument: the accompanying study examines how users interpret, trust and reason with AI-mediated news, and where transparency helps versus where it backfires.

Multi-agent aggregation and cross-checking pipeline
Transparency-oriented validation; multiple framings surfaced side by side
Behavioural user studies on trust and interpretation

Methods: multi-agent systems, LLMs, RAG, behavioural experimentation. Python, Next.js.

ShadowStage · Computational Tholpavakkoothu 2025–26

Tholpavakkoothu is a centuries-old Kerala shadow-puppetry ritual. I treat this cultural form not as inspiration but as a pre-existing computational system: one with rules, constraints and grammars already encoded in its practice. ShadowStage formalises those grammars into a programmable motion system, adds semantic light control, and puts a human performer and a model in the same loop rather than replacing one with the other. Basis of the Computational Tholpavakkoothu paper under review.

Joint-based motion grammars derived from traditional practice
Semantic lighting control tied to dramaturgical state
Human–AI co-performance loop preserving ritual form

Methods: motion systems, generative AI, performance computing. Python, TouchDesigner.

Other work

Narrative Topology Extraction System
Story beats, tension curves, structural transitions and causal density from screenplay datasets.
2024–25
Multimodal Viewer Engagement Analysis System
Facial emotion recognition, blink-rate detection and synchronised engagement curves across a film timeline.
2024–25
DinoSynthesis · Generative Paleo-Acoustics
Hypothetical dinosaur vocalisations via phylogenetic VAE-GAN interpolation. Presented at UAL × CHEAD.
2026
Madhuri · AI Campus Radio
Autonomous LLM and text-to-speech radio with adaptive content sequencing.
2024–25
London Phone-Theft Protest Game
A game used as political argument rather than entertainment.
2026
JOY · Neural Companion System
Privacy-first multimodal companion; processing kept local where the model budget allows.
2025
NODE AI · Decision-Tree Story Generation
Narrative held as a tree of branches that a reader collapses.
2025

08Invited talks & workshops

2026

The Mathematics of Cinema: Patterns, Structure, and Perception in Film

Visiting Lecturer, Creative Computing Institute, University of the Arts London. Computational approaches to cinematic rhythm, audience attention, narrative structure and perceptual modelling.
2026

DinoSynthesis: Generative Paleo-Acoustics via Phylogenetic VAE-GAN

Research Presenter, UAL × CHEAD Generative AI Event, London.
2026

AI and Industry 4.0 Tools for Innovators and Entrepreneurs

Workshop Speaker, Digital University Kerala (online), February 2026.
AI for Humans 101

Invited Speaker, Skill Maaya, ISTE Student Chapter MBCET. AI and machine learning fundamentals for 150+ students from colleges across Kerala.

09Awards & recognition

University of the Arts London Scholarship: £25,000 competitive merit and need-based award for MRes Creative Computing (2025)
Outstanding Student Award, Digital University Kerala (2022–2024 batch)
Runner-Up, Makerversity Under 30s, Somerset House, London (2026)
Second Place, UAL CES × Lovable Portfolio Showcase Gala
First Prize, Smart India Hackathon 2023 (university level)
First Prize, Sharktank Hackathon, Google DSC WoW Kerala
Vidyadhan Scholarship: merit scholarship funded by Infosys co-founder S. D. Shibulal

Technical expertise

AI & computational systems: LLMs, vision-language models, computer vision, multi-agent systems, generative AI, narrative intelligence, RAG, OCR, automation pipelines, behavioural experimentation, saliency modelling, perception analytics
Deep learning & generative modelling: PyTorch, TensorFlow, LangChain, CUDA, GANs, VAEs, audio synthesis, spectral analysis, latent-space modelling
Data science & analytics: Pandas, NumPy, scikit-learn, statistical analysis, EDA, predictive modelling, ML pipelines, A/B testing, hypothesis testing, time-series analysis
Creative AI & media tools: ComfyUI, Stable Diffusion, RunwayML, Midjourney, Kling AI, Pika Labs, Leonardo AI, Adobe Firefly, TouchDesigner, Blender, Three.js, Figma, Webflow, Cargo, Premiere Pro, After Effects
Programming & development: Python, R, React, Next.js, Flask, Git, SQL, API integration, workflow automation, Linux
Visualisation & research tools: Matplotlib, Seaborn, Plotly, Tableau, Power BI, Jupyter, Google Colab, Excel

10Writing

Explanatory notes on statistics and machine learning, published on Medium.

Population and Sample in Statistics Mar 2026
Different Types of Data Mar 2026
Sturge's Rule Mar 2026
What is a Language Model? Mar 2026
Gradient Descent & Learning Rate, for Humans Mar 2026
Auto-Encoders, Simplified for Humans Mar 2026

11Contact

I'm glad to talk about computational perception, narrative structure, AI-mediated media, or collaborations that sit between them.

rahulrajpvr7d.here@gmail.com

CV (PDF) · LinkedIn · GitHub · Medium

I build computational models of how audiences watch films.