AI BriefingHyperconnect news
The latest AI and developer news from Hyperconnect, with short summaries.
Visit website →Part 2: Policy-Following Evaluators, LLM-as-a-Judge
Checklists and critique decomposition pushed the LLM Judge's agreement rate with humans up to 84%.
Part 1: No Data, No Answers: How Hyperconnect Tames LLMs
The problem of generating explanations with no ground-truth data is solved through policy consensus and iterative evaluation.
On-device AI Face Verification Pipeline Optimization
On-device Android face verification optimization improved latency by 37% and throughput by 530%.
How to Relax Business Problems into AI Problems
Business goals must be relaxed into smaller AI optimization problems to create impact.
Why Doesn't Deployment Work? Designing ML Model Training and Evaluation to Match Target Metrics
We redesigned training and evaluation metrics to align with the goal of directly increasing conversion rate.
Start Experiments with One Click! Building a Recommendation Experiment System That Shortens Iteration Cycles
Azar used PolicyEngine to cut experiment setup and review time from 2 days to 30 minutes.
How Does a 1:1 Video Chat Service Automate E2E Regression Testing
By automating 308 tests across 4 platform combinations, the time was reduced from 102 hours to 7 hours.
Building a Flink-Based Stream Join Service for AI Real-Time Recommendation Systems
Using Flink's KeyedProcessFunction, they implemented real-time event combination and zero-downtime deployment.
Building a robust, high-performance ScyllaDB cluster with Super Disk
Super Disk's write-mostly RAID cut ScyllaDB recovery time by up to 10x.
Finding End-to-End Latency Bottlenecks in Apache Flink Applications
A method to narrow down Flink latency bottlenecks by separating processing time from non-processing time.
Adopting Flink SQL: Implementation and Operational Experience
They replaced a legacy Flink app running on 96 CPUs with Flink SQL and GitOps deployment.
A Deep Dive into Spring Transactional Rollback
Proxies, rollback marking, propagation options, and Kotlin exception conversion determine what actually gets rolled back.