View PDF ↗
PDF Viewer

Loading PDF...

This may take a moment

BUILDER'S SANDBOX

Core Pattern

AI-generated implementation pattern based on this paper's core methodology.

Implementation pattern included in full analysis above.

MVP Investment

$9K - $12K
6-10 weeks
Engineering
$8,000
Cloud Hosting
$240
SaaS Stack
$300
Domain & Legal
$100

6mo ROI

2-4x

3yr ROI

10-20x

Lightweight AI tools can reach profitability quickly. At $500/mo average contract, 20 customers = $10K MRR by 6mo, 200+ by 3yr.

Talent Scout

G

Gabriele Conte

Politecnico di Bari

A

Alessio Mattiace

Politecnico di Bari

G

Gianni Carmosino

Politecnico di Bari

P

Potito Aghilar

Politecnico di Bari

Find Similar Experts

Personal experts on LinkedIn & GitHub

Founder's Pitch

"RUVA offers on-device, transparent, and editable personal AI knowledge management, ensuring user privacy and control."

Personal AIScore: 9View PDF ↗

Commercial Viability Breakdown

0-10 scale

High Potential

2/4 signals

5

Quick Build

4/4 signals

10

Series A Potential

4/4 signals

10

🔭 Research Neighborhood

Generating constellation...

~3-8 seconds

Why It Matters

RUVA addresses critical privacy and security gaps in existing AI systems by giving users full control over their data and its usage, thereby ensuring transparency and the right to be forgotten.

Product Angle

RUVA can be productized as a mobile application providing users with full control over their personal data, ensuring transparency and privacy, crucial for PDPR compliance and appealing to privacy-focused consumers.

Disruption

RUVA could replace existing personal data management tools and RAG systems that do not allow detailed user interaction with the data.

Product Opportunity

The demand for enhanced digital privacy tools is significant, especially in light of regulatory developments like GDPR and CCPA. Consumers and enterprises are willing to pay for solutions that offer transparency and control over data.

Use Case Idea

A mobile app for privacy-conscious users to manage their personal data, such as emails and photos, in a transparent and editable way.

Science

RUVA utilizes a neuro-symbolic GraphRAG architecture to create a Personal Knowledge Graph. This allows on-device graph reasoning instead of vector matching for data retrieval, making it possible for users to inspect, edit, and delete specific nodes and edges in their data representation.

Method & Eval

The solution is tested on Google Pixel 8 Pro with interactive latencies under mobile constraints and achieves high semantic accuracy and reasoning capability tested against a benchmark of diverse data queries.

Caveats

The requirement for local device operation limits compute-heavy tasks. Users may need technical knowledge to effectively manage their data, and device storage constraints could affect scalability.

Author Intelligence

Gabriele Conte

Politecnico di Bari
g.conte12@studenti.poliba.it

Alessio Mattiace

Politecnico di Bari
a.mattiace@studenti.poliba.it

Gianni Carmosino

Politecnico di Bari
g.carmosino1@studenti.poliba.it

Potito Aghilar

Politecnico di Bari
potito.aghilar@poliba.it

Giovanni Servedio

Politecnico di Bari
giovanni.servedio@poliba.it

Francesco Musicco

Politecnico di Bari
f.musicco@phd.poliba.it

Vito Walter Anelli

Politecnico di Bari
vitowalter.anelli@poliba.it

Tommaso Di Noia

Politecnico di Bari
tommaso.dinoia@poliba.it

Francesco Maria Donini

Università degli Studi della Tuscia
donini@unitus.it

References (16)

[1]
Qwen3 Technical Report
2025An Yang, Anfeng Li et al.
[2]
The widespread adoption of large language model-assisted writing across society
2025Weixin Liang, Yaohui Zhang et al.
[3]
When Machine Unlearning Meets Retrieval-Augmented Generation (RAG): Keep Secret or Forget Knowledge?
2024Shang Wang, Tianqing Zhu et al.
[4]
From Local to Global: A Graph RAG Approach to Query-Focused Summarization
2024Darren Edge, Ha Trinh et al.
[5]
CheckEval: A reliable LLM-as-a-Judge framework for evaluating text generation using checklists
2024Yukyung Lee, Joonghoon Kim et al.
[6]
Prometheus: Inducing Fine-grained Evaluation Capability in Language Models
2023Seungone Kim, Jamin Shin et al.
[7]
An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-Tuning
2023Yun Luo, Zhen Yang et al.
[8]
An Ecosystem for Personal Knowledge Graphs: A Survey and Research Roadmap
2023M. Skjæveland, K. Balog et al.
[9]
The Third AI Summer: AAAI Robert S. Engelmore Memorial Lecture
2022Henry A. Kautz
[10]
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
2020Patrick Lewis, Ethan Perez et al.
[11]
Personal Knowledge Graphs: A Research Agenda
2019K. Balog, Tom Kenter
[12]
From Louvain to Leiden: guaranteeing well-connected communities
2018V. Traag, L. Waltman et al.
[13]
The proof and measurement of association between two things.
2015C. Spearman
[14]
Interrater reliability: the kappa statistic
2012M. McHugh
[15]
Computing Krippendorff's Alpha-Reliability
2011K. Krippendorff
[16]
Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit.
1968Jacob Cohen