Project
RevFlōw
Intelligent transaction categorisation and financial coaching for Revolut
A financial intelligence platform built in 24 hours at HackDelft 2026 for the Revolut challenge. It cuts through "The Data Fog" — broken merchant labels and generic categorisation — with a two-stage ML pipeline (rule pre-pass + CalibratedLinearSVC on sentence embeddings), KMeans life-stage personalisation, an Isolation Forest subscription forensic engine, and a Claude Haiku money coach with one-tap action cards.
RevFlōw is a financial intelligence platform built during the Revolut challenge at HackDelft 2026 in 24 hours with Team YUBRE. It addresses "The Data Fog" — the trust breakdown caused by generic merchant labels and failed categorisation — with a four-stage intelligence pipeline.
Intelligent categorisation
Keyword pre-pass with 18 Dutch merchant rules (confidence = 1.0 short-circuits the ML stage)
CalibratedLinearSVC trained on MiniLM-L6 384-d sentence embeddings + OHE features (entry method, merchant domain, amount) across 12 spending categories
75% accuracy on high-confidence predictions (≥0.45); uncertain merchants fall back to "Miscellaneous" rather than producing wrong labels
<50 ms inference latency end-to-end
Personalisation
KMeans clustering segments users into four life-stage cohorts: Student, Family, Professional, Senior
Per-category spend share re-weights global model probabilities (
P' ∝ P_global × f_user) so categorisation fits the user's actual life stageEach categorised transaction is returned with a human-readable explanation grounded in user history
Subscription insight engine
Six parallel signal detectors: 4-week z-score anomalies, CV + cadence subscription detection, new merchant bursts, fraud flags (velocity, smurfing), cohort vs. peers comparison, and Isolation Forest user profiling
Detects forgotten, overlapping, or creeping recurring charges before they compound into "subscription drag"
Emits structured insight objects (type · severity · metrics) that drive both the deterministic action cards and the AI coach
AI Money Coach & Command Centre
Insights feed Claude Haiku for concise two-sentence advice; supplier-agnostic, self-hosted path available in production
Deterministic action cards for 1-tap interventions: spending limits, recurring charge cancellation, analytics drills
Revolut-styled Command Centre with transaction timeline, monthly budget ring, live insight cards, and voice query input
Deployed to Railway; live at revflow.brewen.dev
A financial intelligence platform built in 24 hours at HackDelft 2026 for the Revolut challenge. It cuts through "The Data Fog" — broken merchant labels and generic categorisation — with a two-stage ML pipeline (rule pre-pass + CalibratedLinearSVC on sentence embeddings), KMeans life-stage personalisation, an Isolation Forest subscription forensic engine, and a Claude Haiku money coach with one-tap action cards.
RevFlōw is a financial intelligence platform built during the Revolut challenge at HackDelft 2026 in 24 hours with Team YUBRE. It addresses "The Data Fog" — the trust breakdown caused by generic merchant labels and failed categorisation — with a four-stage intelligence pipeline.
Intelligent categorisation
Keyword pre-pass with 18 Dutch merchant rules (confidence = 1.0 short-circuits the ML stage)
CalibratedLinearSVC trained on MiniLM-L6 384-d sentence embeddings + OHE features (entry method, merchant domain, amount) across 12 spending categories
75% accuracy on high-confidence predictions (≥0.45); uncertain merchants fall back to "Miscellaneous" rather than producing wrong labels
<50 ms inference latency end-to-end
Personalisation
KMeans clustering segments users into four life-stage cohorts: Student, Family, Professional, Senior
Per-category spend share re-weights global model probabilities (
P' ∝ P_global × f_user) so categorisation fits the user's actual life stageEach categorised transaction is returned with a human-readable explanation grounded in user history
Subscription insight engine
Six parallel signal detectors: 4-week z-score anomalies, CV + cadence subscription detection, new merchant bursts, fraud flags (velocity, smurfing), cohort vs. peers comparison, and Isolation Forest user profiling
Detects forgotten, overlapping, or creeping recurring charges before they compound into "subscription drag"
Emits structured insight objects (type · severity · metrics) that drive both the deterministic action cards and the AI coach
AI Money Coach & Command Centre
Insights feed Claude Haiku for concise two-sentence advice; supplier-agnostic, self-hosted path available in production
Deterministic action cards for 1-tap interventions: spending limits, recurring charge cancellation, analytics drills
Revolut-styled Command Centre with transaction timeline, monthly budget ring, live insight cards, and voice query input
Deployed to Railway; live at revflow.brewen.dev
References
Transaction Categorization with Relational Deep Learning in QuickBooks
Dong, K., Jonnalagedda, P., Gao, X., Acharya, A., Kissa, M., Flores, M., Chawla, N. V., Das, K.. (2025). arXiv preprint arXiv:2506.09234. DOI: 10.48550/arXiv.2506.09234
Categorising SME Bank Transactions with Machine Learning and Synthetic Data Generation
Aluffi, P. A., Jess, B., Bazzi, M., Kennedy, K., Arderne, M., Rodrigues, D., Lotz, M.. (2025). arXiv preprint arXiv:2508.05425. DOI: 10.48550/arXiv.2508.05425
Using Zero-shot Prompting in the Automatic Creation and Expansion of Topic Taxonomies for Tagging Retail Banking Transactions
de S. Moraes, D., Santos, P. T. C., Costa, P. B. D., Pinto, M. A. S., Pinto, I., Veiga, Á., Colcher, S., Busson, A., Rocha, R. H., Gaio, R., Miceli, R., Tourinho, G., Rabaioli, M., Santos, L., Marques, F., Favaro, D.. (2024). arXiv preprint arXiv:2401.06790. DOI: 10.48550/arXiv.2401.06790
Enhancing Foundation Models in Transaction Understanding with LLM-based Sentence Embeddings
Fan, X., Jiang, Z., Yeh, C.-C. M., Chen, Y., Dou, Y., Pan, M., Zheng, Y.. (2025). arXiv preprint arXiv:2601.05271. DOI: 10.48550/arXiv.2601.05271
Hierarchical Classification of Financial Transactions Through Context-Fusion of Transformer-based Embeddings and Taxonomy-aware Attention Layer
Busson, A., Rocha, R. H., Gaio, R., Miceli, R., Pereira, I., Moraes, D. D. S., Colcher, S., Veiga, Á., Rizzi, B., Evangelista, F., Santos, L., Marques, F., Rabaioli, M., Feldberg, D., Mattos, D., Pasqua, J., Dias, D. G.. (2023). arXiv preprint arXiv:2312.07730. DOI: 10.48550/arXiv.2312.07730
Clustering in Recurrent Neural Networks for Micro-Segmentation using Spending Personality
Maree, C., Omlin, C.. (2021). arXiv preprint arXiv:2109.09425. DOI: 10.48550/arXiv.2109.09425
Your Spending Needs Attention: Modeling Financial Habits with Transformers
Braithwaite, D. T., Cavalcanti, M., McEver, R., Udagawa, H., Silva, D., Ramanath, R., Meneses, F., Yoshida, A., Wingert, E., Ramos, M., Zanfelice, B., Gupta, A.. (2025). arXiv preprint arXiv:2507.23267. DOI: 10.48550/arXiv.2507.23267
Open Banking Foundational Model: Learning Language Representations from Few Financial Transactions
Polleti, G., Santana, M., Fontes, E. R.. (2025). arXiv preprint arXiv:2511.12154. DOI: 10.48550/arXiv.2511.12154
FlowSeries: Anomaly Detection in Financial Transaction Flows
Capozzi, A., Vilella, S., Moncalvo, D., Fornasiero, M., Ricci, V., Ronchiadin, S., Ruffo, G.. (2025). arXiv preprint arXiv:2503.15896. DOI: 10.48550/arXiv.2503.15896
Advancing Anomaly Detection: Non-Semantic Financial Data Encoding With Large Language Models
Bakumenko, A., Hlaváčková-Schindler, K., Plant, C., Hubig, N. C.. (2024). arXiv preprint arXiv:2406.03614. DOI: 10.48550/arXiv.2406.03614
Enhancing Anomaly Detection in Financial Markets with an LLM-based Multi-Agent Framework
Park, T.. (2024). arXiv preprint arXiv:2403.19735. DOI: 10.48550/arXiv.2403.19735
Anomaly Detection in High-Dimensional Bank Account Balances via Robust Methods
Maddanu, F., Proietti, T., Crupi, R.. (2025). arXiv preprint arXiv:2511.11143. DOI: 10.48550/arXiv.2511.11143
Synthesizing Behaviorally-Grounded Reasoning Chains: A Data-Generation Framework for Personal Finance LLMs
Theerthala, A.. (2025). arXiv preprint arXiv:2509.14180. DOI: 10.48550/arXiv.2509.14180
Conv-FinRe: A Conversational and Longitudinal Benchmark for Utility-Grounded Financial Recommendation
Wang, Y., Han, Y., Qian, L., He, Y., Peng, X., Feng, D., Xie, Z., Zhang, V., Guo, R., Mo, F., Huang, J., Chen, Y., Liu, X., Nie, J.-Y.. (2026). arXiv preprint arXiv:2602.16990. DOI: 10.48550/arXiv.2602.16990