I built a lead-intent pipeline on 1.75M posts/day for €10/day. It got 1 trialOn September 15, I stopped Dwelverson. TL;DR Seven months solo on Dwelverson, a B2B lead-intent product: read ~1.75M social posts a day, find the ones where someone is looking to buy what a client se16MSep 21 · 27 min readvivian-oliveres.hashnode.dev
What Is Lead Intent Detection, and Why I'm Betting on ItAfter my V2 digest failed commercially, I knew I needed to pivot to B2B. The technology worked. The market didn't. But before building anything new, I had to understand the domain I was entering. I spMar 22 · 11 min readvivian-oliveres.hashnode.dev
How My Bluesky Digest (V2) MVP FailedMy previous post ended with a promise: I was going to rebuild the ranking from scratch. Train a proper ML model. Use BM25 and embeddings as features, not as the final answer. And use LLMs to generate Mar 13 · 12 min readvivian-oliveres.hashnode.dev
How My Bluesky Digest MVP Failed, and What I Learned Building ItI spend a lot of time reading technical content on social media. Articles, threads, expert takes. And once in a while, I stumble on something rare: a short exchange between two (or even more) experts Mar 3 · 8 min readvivian-oliveres.hashnode.dev
20 Years of Code, Zero Years of Product: Why I'm Building in PublicI've been writing code for nearly 20 years. Backend systems, data pipelines, ML tooling, infrastructure... Across banks, various startups, and big tech companies. Always for someone else's. Now I'm buFeb 27 · 3 min readvivian-oliveres.hashnode.dev