Eban Schachter@eschaq
3
Events attended
3
Submissions made
1 year of experience
About me
Marketing Analytics Manager with 30 years of enterprise experience and an AI builder closing the gap between data and decisions. I create tools that help businesses turn data and market signals into clear, actionable intelligence — combining real-world analytics expertise with modern AI to help teams make faster, smarter decisions.
🤓 Latest Submissions
Plainlode: The Signal, Mined Plain
A solo online store owner has to decide what to sell next from live market signal scattered across a dozen tools, with no analyst and no time. So they guess. Plainlode ends the guessing. You name any category. Plainlode finds the products worth watching, then reads live demand (Google Trends) and live supply (Amazon product data) on each. It ranks what is actually moving and hands you one plain-language call: findings, options, and a recommended action. The recommendation reasons over both demand and supply. When demand is rising but the market is saturated, Plainlode does not say "stock more," it says differentiate. And it does something most tools never will: it argues against its own call, naming the single live signal that would reverse the recommendation. A call you can trust is one that tells you when to stop trusting it. Under the hood, two model steps run on AMD Instinct through Fireworks AI: a cheap-tier filter that reads the signal, and a plain-language engine that writes the briefing. Facts come from live retrieval, never training, so the answer is never stale. If a source stumbles, it falls back to recent real data and never fabricates. Live today at plainlode-production.up.railway.app. The demo is the wedge. The engine is the company.
13 Jul 2026
Foreshock — Continuous ICT Vendor Risk Monitoring
Under DORA Article 28, fintechs are legally accountable for continuous oversight of every critical ICT vendor. The leading indicators of a vendor going bad show up in public data weeks before they reach a security score or a questionnaire cycle. Leadership exits. Lawsuits. Hiring freezes. Sentiment collapse. GRC platforms watch paperwork. Security raters watch the attack surface. Neither one watches business health, and that is the gap Foreshock fills. Foreshock runs a daily unattended agent. It pulls signals across five query classes per vendor through Bright Data MCP, and for public companies it pulls SEC EDGAR 8-K filings straight from the source. Every signal gets appended to a Type-2 timestamped history and never overwritten, so the trend is always preserved. A Claude validator throws out the false positives (about 80% of candidates), and a CDC diff scores six weighted dimensions: leadership, legal, headcount, sentiment, news volume, and open roles. When several signals deteriorate at once, a convergence alert fires. AI then writes the risk summary the way a GRC analyst would, and every factual claim carries a citation that resolves to its source signal. A built-in citation audit confirms it, with zero unresolved across all vendors. One click exports a DORA Article 28 ICT Register PDF: cover page, fleet audit, per-vendor sections, the AI narrative with numbered sources, and a methodology appendix. No competitor ships that today. The same engine (watch, detect, score, alert, summarize, source) points at any entity where stale data defeats the purpose. Fintech vendors today. Competitors, suppliers, and acquisition targets next.
31 May 2026
Prism — Intelligence for Every Audience
Your customers are telling you what they need. They post it on Reddit, talk about it on podcasts, leave it in reviews, vent it on Hacker News. Your internal data is telling you what's actually happening in the business. Both signals are loud, and almost nobody is reading them together. That gap is where decisions stall. Marketing pitches a campaign that ignores what customers are actually complaining about. Finance builds a forecast against a market that just shifted. The analyst who could connect the dots is busy rewriting the same insight four different ways for four different audiences. Prism closes the gap. It ingests external signals (Reddit, RSS, podcasts, Hacker News) alongside your internal data (CSV, Excel, anything you upload), scores what matters using our Pain Point Severity framework, and writes the briefing each audience needs. A CFO gets the financial case. An Ops lead gets the operational impact. Marketing gets the message. Sales gets the objection map. Thirteen audience profiles in all, each fed by the same underlying intelligence. Built on React, FastAPI, and Claude. Live now on HuggingFace Spaces.
19 May 2026
