“I don't want to be the go-between. I'm busy. I want a three-way conversation between you, the other party, and me.”
David Joerg
Technical Product Manager, AI/ML at Chess.com
How a tech entrepreneur turned an AI coworker into a full operating system for work and life. Then his family did the same.
David Joerg
Technical Product Manager, AI/ML at Chess.com
“I don't want to be the go-between. I'm busy. I want a three-way conversation between you, the other party, and me.”
David Joerg
Technical Product Manager, AI/ML at Chess.com
6
weeks of usage
30+
distinct projects
6
family & friends onboarded
24/7
autonomous tasks running
David Joerg is a software developer and entrepreneur based in New York City. He co-founded Vindigo, one of the first popular mobile apps in the US. He was an early team member at Two Sigma Investments, now a leading quantitative investment firm with $50B+ AUM.
Today, David is Technical Product Manager, AI/ML at Chess.com, where he created "Learn Chess with Dr. Wolf," an AI chess coaching app with 5M+ downloads. He now builds AI-powered coaching experiences at Chess.com for its 250M+ users.
When he started using Viktor in late February 2026, he didn't need hand-holding. He needed a coworker who could keep up.
Across 6 weeks, David and Viktor tackled over 30 projects spanning software engineering, real-world logistics, research, creative production, and financial analysis. Here are the top 10 by effort and impact:
Built a platform for making and receiving live AI phone calls with scripted multi-act conversations, a human-approval queue, and voice auditions across 10 candidate voices. Used for cold outreach, restaurant reservations, and professional interviews.
Built and daily-evolved an interactive web app over 12 iterations with GitHub CI/CD, deployed to a custom domain. Each morning brought a new version: feature additions, visual overhauls, and design pivots, all shipped directly to production.
Set up SMS-to-Slack routing via Twilio, a custom email domain with webhook-based receiving and junk filtering, and the AI phone-calling capability. Built from scratch, this became the backbone for many other projects.
What would have taken 1 to 2 weeks of dedicated full-time research now runs autonomously every day without being asked. Viktor built and maintains a 70+ entry database in Notion via daily automated scans of forums, medical directories, and Reddit.
Each trip would have taken days to plan. Viktor handled everything end-to-end: event tickets, lodging, rail connections, restaurant reservations for specific dietary requirements, real-time airport terminal monitoring, and day-by-day itineraries delivered as PDFs.
A fact-checked advisory board slide deck across 4 major versions, each with separate passes for fact-checking, audience review, and complete rewrites. Every claim was backed by verifiable sources with real URLs.
Conducting live AI-powered phone interviews with business professionals about their AI usage, then synthesizing findings into a group presentation. Viktor scripts the calls, makes them, and analyzes the results.
Processed 8+ invoices and legal documents to produce an adjusted cost basis calculation for a property transfer. Required careful reading of contracts, amendment tracking, and LLC accounting. The thoroughness directly translates into major future tax savings: receipts and records that a manual pass would have missed.
Found a new garage with over 40% cost savings after researching 7 garages, making 10+ AI phone calls across 4 rounds (using two different voice AI platforms), and running a multi-week email campaign. David never made a single call himself.
Built a proxy-based scraper that cataloged 209 videos with AI-generated analysis, delivered as a comprehensive spreadsheet for content strategy review.
What started as one user quickly became a team. David's family and friends adopted Viktor independently, without onboarding docs or training sessions. They just saw what it could do and started asking.
Became an independent power user for international travel planning (day-by-day itineraries), restaurant bookings (vegetarian-focused research), fashion lookbooks, financial analysis of loans and property sales, LLC dissolution review across 15 legal documents, and settlement structuring. All on her own initiative.
Became an independent power user for international travel planning (day-by-day itineraries), restaurant bookings (vegetarian-focused research), fashion lookbooks, financial analysis of loans and property sales, LLC dissolution review across 15 legal documents, and settlement structuring. All on her own initiative.
Uses Viktor for interpreting medical imaging results, tax predictions, and event planning. Went from 28 days of silence after initial DM to daily active user once he saw the value.
Saw what David and Family Member 2 were doing, then set up his own independent Viktor account. Within a weekend, he and a friend who owns a car dealership built a working "Deal Scanner" that finds high-margin private-party deals on listing sites, analyzes them against current market values, generates AI pricing and talking points, and auto-drafts buyer outreach messages. Viktor produced the cost analysis, breakeven model, and the entire proof of concept. Live and functional.
"Find me a parking garage" means: research options, call them, compare prices, email the best ones, negotiate, and report back with a recommendation. "Plan the dinner party" means: design the menu, place the catering order, and create a custom movie poster for the invitation. End-to-end execution, not suggestions.
The real impact isn't time saved — it's the projects that would never have happened. Without Viktor, David wouldn't have built a 70-entry medical research database (no time), would have declined the advisory board opportunity or shown up underprepared (no bandwidth), and would have skipped the multi-city trip planning or accepted a worse trip.
Most projects involved writing and deploying real software: Python scripts, web applications with CI/CD pipelines, Twilio integrations, browser automation, proxy-based web scrapers, Notion API integrations, ElevenLabs voice synthesis. Viktor doesn't just describe what could be built — it builds, deploys, and iterates on it daily.
Viktor makes phone calls to real businesses and real people. It sends emails from a professional custom domain. It texts via SMS. It books restaurant reservations. It orders products for delivery. It monitors flight terminals in real time.
Multiple workstreams run continuously without human prompting. The medical practitioner database scans the internet daily and adds new entries to Notion. Morning briefings arrive every day. The web app evolved through 12+ daily iterations.
David is a rigorous thinker. He challenges claims, demands real source URLs, insists on unassailable arguments, and pushes back on generic advice. The advisory board presentation went through 4 major versions with separate passes for fact-checking, audience analysis, and complete rewrites. Viktor adapted to this standard and maintained it across every project.
>40%
parking cost reduction
209
videos cataloged & analyzed
70+
database entries (autonomous)
12
daily app iterations
In 6 weeks, Viktor went from a new tool in David's Slack to the operating system for his professional and personal projects. It writes code, makes phone calls, sends emails, builds databases, plans trips, analyzes legal documents, and creates media — all from Slack messages.
More importantly, it scales. When a family member needed an international travel itinerary, she didn't ask David — she asked Viktor. When another family member needed to conduct professional interviews, he didn't hire a research firm — he used Viktor. When a friend needed a medical imaging report interpreted, he asked Viktor.
“The bigger point is that without Viktor I would have never done these things. I don't have the time, nor do I have the time to recruit and supervise someone to do these tasks to the level of quality that Viktor achieves.”
David Joerg
Technical Product Manager, AI/ML at Chess.com
Use Cases
TWL
Australian ecommerce retailer
Replacing ~2 hours of manual work per day, split across 5 team members at an Australian functional-fitness retailer.
Element Turf
Lawn care and landscaping company
How a 25-person, 8-crew landscaping company put an AI coworker to work across Aspire, ClickUp, BambooHR, Gmail, and seven inboxes.
David started with one Slack message. Six weeks later, Viktor runs 30+ projects for his entire family.