The Jewish work
2004 – 2008
Young Judaea, San Juan
Vice-president of a 200-member Jewish youth movement through high school in Puerto Rico, where I was born and raised - running activities, budgets, and fundraising with a 15-person board. Summers at a Jewish heritage camp in North Carolina; Israel in 2006.
The engineering career
2004 – 2005
First paid technical work
Building and programming custom computers for customers in San Juan, at fourteen.
The engineering career
2008 – 2013
Cornell University
B.S. in Information Science, Systems & Technology; M.Eng in Engineering Management with a computer science concentration. Taught calculus as a TA. Treasurer of the Society of Hispanic Professional Engineers.
The engineering career
Summers 2010 · 2011 · 2012
Goldman Sachs - three internships
Software engineering intern on the same trade-processing team three summers running - building tools that double-check the firm's trade records against its in-house risk system.
The engineering career
May 2013 – June 2015
Goldman Sachs - full time
Software engineer in Securities technology - the systems that double-check trades and file the firm's regulatory reports. I built a data-lookup library used across the firm and helped stand up a round-the-clock operation that catches and fixes trade errors.
The Jewish work
2013
The collecting begins
A few months into Goldman, I started saving online Torah learning to disk - recorded classes (shiurim), essays, short teachings. The archive never stopped growing. Years later it became training data.
The Jewish work
c. 2014
Yeshiva before work
5:50 to 8:30 every morning learning at yeshiva, then the trading floor. The résumé I wrote that year lists volunteer Jewish-outreach work with MEOR and Chabad - and a goal I'd already put into words: improving Jewish learning “on and offline.”
June 2015
Working for myself, under two names
I left Goldman and founded two companies in the same month: LazoffTech, a software engineering practice, and TorahTech LLC. From here on, the consulting funded the Torah work. It still does.
The Jewish work
2016 – 2018
Paid to build Jewish education
Senior engineer on Project Mercava, a learning platform built with a team in Israel and tailored to the Orthodox Jewish learning style; then a comprehensive redesign of Aleph Beta's Torah video platform. The niche existed - I was working in it.
TORAH.RTF · JANUARY 5, 2018 · ONE MONTH BEFORE MY FIRST DAY IN THE BLOCKCHAIN INDUSTRY
“These books look like any other. But they are actually the records of our sages, generation after generation on the same source, the Torah. It's like the blockchain of history, it's been copied exactly, generation to generation, with very specific rules.”
I then spent six years learning the technology that could test that idea.
The engineering career
February 2018 – February 2024
ConsenSys - six years
I joined through ConsenSys Academy, building its online learning platform, and stayed six years - ending as the senior engineering manager for every website behind metamask.io and consensys.io: more than 100 sites serving millions of users.
The Jewish work
2019
Campus outreach, nights and weekends
Alongside the day job: web copy, program pages, and ad campaigns for MEOR, a Jewish campus-outreach organization - including a campaign aimed at my own alma mater.
The engineering career
May 2021 – October 2024
Chainlink Labs - three and a half years
Contract engineer inside Chainlink Labs - a company that feeds real-world data like prices and events onto blockchains - building the websites and tools its developers rely on. The contract ran through TorahTech LLC: the blockchain income flowed through the Torah company.
The engineering career
February 2023
Digital Twin Pro founded
Everything the career taught - connecting information, making it visual, running systems that don't lose things - aimed at one product: private AI on hardware you own.
The Jewish work
Now · 2026
OpenRabbi
The archive that started in 2013, rebuilt as a knowledge graph: 17,147 rabbinic teachings, each carrying its Hebrew source and a citable reference, linked to 33,323 verses inside a 157,143-node graph (measured August 2026) - grounding AI voices that cite their sources or say nothing at all.
The test came back
Did the idea hold up?
It held, and it bit. At one point I fed the AI roughly 134,000 extra texts with no citations attached. More material, better answers, right? The answers got measurably worse. The graph had named Rashi 3,017 times and stored not one sentence of his actual words. So I enforced the sages' own rule in code: report a teaching in the name of the one who said it, or say nothing. That rule is what made the machine honest.
The discipline that makes a machine cite a 900-year-old commentary, or admit it doesn't know, is the same discipline inside the box I now ship to customers. That box is Digital Twin Pro.