When we decided to take the Q2 @solanaturbine Advanced SVM program in a slightly different direction, we knew we were setting ourselves a difficult challenge. Training engineers on the Solana application layer is hard enough. Helping them transition to the protocol layer in just six weeks is an entirely different challenge. But we believed the demand for protocol engineers will only continue to grow as the SVM ecosystem matures. Over the last three weeks, our cadets have spent countless hours diving deep into the SVM, networking, QUIC, Alpenglow, and Constellation. We've also had engineers working directly on the protocol share their experiences, design tradeoffs, and current challenges. The exciting part is that our cadets didn't stop at theory. In just a few weeks, they've already shipped practical systems experiments around Alpenglow, MCP, packet processing, and banking-stage performance: • xdp-pshred-forwarder- implementing an XDP demultiplexer against the pshred wire format and proposer identity field • svm-xdp-packet-router- exploring kernel-level packet routing strategies for SVM traffic flows • banking-stage-jitter-harness experiments- a measurement harness for evaluating banking-stage timing behavior against Alpenglow Votor round requirements A few examples: github.com/ChiefWoods/xdp-ps… github.com/aspnxdd/svm-xdp-p… github.com/Allen-Saji/bankin… github.com/jkrishnad/banking… github.com/SergZen/agave/tre… This is only a small sample of what they've built in the first few weeks. Watching engineers go from application development to building protocol-level systems in such a short period has been incredibly rewarding. And we're only getting started. The next few weeks will involve even more experimentation on the Read layer, Alpenglow, QUIC-only ingestion, MCP, profiling, benchmarking, and validator performance engineering. Plenty of glass-chewing ahead.