Company
Ramp
12 cases from Ramp (Fintech), each digested from a public talk or engineering post and linked to the original. 1 was checked in an online A/B test; 8 report results in production.
Tasks: Evaluate models, Optimize decisions, Match entities, Detect anomalies, Catch fraud and abuse and 5 more. Methods: Golden set / benchmark, Human in the loop, LLM classification / extraction, Embedding similarity matching, OCR, Thompson sampling and 33 more. Tools: NumPy, Gemini, Claude, XGBoost, Ray, PyTorch and 11 more.
Tech blog: engineering.ramp.com
| Agentic Credit Analysis | fintech | 2026-10 | |
| Putting a Transformer Inside Every Card Authorization | fintech | 2026-10 | |
| Online Learning for Cost-Efficient LLM Routing | fintech | 2026-07 | |
| Agentic Risk Operations | fintech | 2026-06 | |
| We Tested Marketing Incentives to AI Agents. Here's What Happened. | fintech | 2026-04 | |
| Automating Receipt Collection: Apple Intelligence for On-Device Inference | fintech | 2026-04 | |
| Financial Benchmarks | fintech | 2026-03 | |
| How To Build Agents Users Can Trust | fintech | 2025-07 | |
| How Ramp Fixes Merchant Matches with AI | fintech | 2025-06 | |
| From RAG to Richness: How Ramp Revamped Industry Classification | fintech | 2025-01 | |
| Improving Retrieval on Ramp with Transaction Embeddings | fintech | 2024-08 | |
| Make Better Decisions by Embracing Uncertainty | fintech | 2024-06 |