As someone walks me through a problem, I can visually picture the whole system in my head: the tools, the handoffs, the data moving between them.
That’s how I’m able to ship fast and find inefficiencies. The picture becomes clear. I just need to bring it to life.
02 the angle
Then I see the optimal path through.
Most people don’t see the full map. They don’t see all the tools available to solve the problem. And they don’t see the optimal route between them.
I do that in my head. Once I’m holding the whole picture, I can think through unique and unconventional ways to solve it. Routes that only surface when you can see every component and every join at once.
Most of the work happens here. By the time anything gets built, the hard part is already done.
03 the build
Then I ship it.
Once the route is clear, the build is the easy part. Agents, automations, data analysis, reports, custom tools. Whatever shape the solution takes, it gets wired up and goes live.
I move fast because the design is already done. Most of the work happened in my head. This part is just bringing it to life.
04 built
Lately.
A few of the things I’ve been working on. Different domains, same instinct. Each one is the same three steps applied to a different system. See it, find the angle, build it.
// shippingv0.4 · 4d ago
Control Labs
A chief-of-staff agent that consolidates Slack, Linear, Notion, email and your own head into one delegatable pipeline. Action items stop falling through.
// shipping · betav0.2 · 11d ago
Orvo.health
Hospital pricing transparency. Compare a bill against self-pay, insurance and Medicare rates and walk in armed to negotiate.
// shipped
EdgarIQ
Multi-LLM RAG over the entire SEC filings corpus. Ask a question of every 10-K at once. Citations baked in.
// shipped
LegacyHash
A non-custodial inheritance protocol for crypto. Wallets that survive their owner without surrendering custody during their life.
Background in quantitative research and investment banking. A decade working with data, models, and complex systems across defense, banking, crypto, and AI.
Today I build agentic systems. AI workflows, tools, and automation that operate like a scalable workforce.
0-to-1 on the modern stack. Python, SQL, TypeScript, Next.js, Postgres. Research to production.
05.1 waterloo
Biotech and economics.
University of Waterloo · BSc Honours, Biotechnology & Economics
Honours BSc in Biotechnology and Economics. Biotechnology: decomposing complex biological systems into testable components and tracing cause through them. Economics and econometrics: regression, time-series, causal inference, statistical work with messy real-world data.
05.2 idf
Sergeant First Class.
Israel Defense Forces · [unit redacted]
Led 120+ soldiers through 50+ missions, with advanced training in urban warfare and counter-terrorism. Developed strong management, leadership, teamwork, and communication skills, with real-life experience problem-solving under acute pressure.
05.3 bloom_burton
Healthcare banking.
Bloom Burton & Co. · Toronto · IB Analyst
Origination and execution of equity financings, M&A advisory, and capital markets transactions in the healthcare sector, collaborating with senior bankers, client teams, and investors on stringent deadlines.
Built complex financial and valuation models using DCF, working with internal and client teams to define operating assumptions, comparable companies, and precedent transactions to price an IPO and support preliminary M&A advisory.
Performed detailed financial statement analysis across filings, investor presentations, and research reports to benchmark multiples, evaluate capital structure options, and provide analytical support.
05.4 albert_street_capital
DeFi fund. Co-founder.
Albert Street Capital · Chicago · Co-Founder, Managing Director
Founded and run a cryptocurrency fund focused on decentralized finance and cross-chain arbitrage. A multifaceted investment strategy: delta-neutral spot/perp, low-risk yield farming and staking, cross-chain arbitrage, and selective positions in DeFi projects with utility tokens.
Built, programmed, and automated the fund’s backend infrastructure end to end. Cryptocurrency portfolio management, dashboards, optimized token tracking, arbitrage execution, and fund administration.
05.5 pine_grove_ventures
Personal builds.
Pine Grove Ventures · Founder, Builder
Personal venture studio for scoping, prototyping, and shipping 0-to-1 products on nights and weekends. Collaborate with AI to compress time-to-MVP from months to hours, and productionization from quarters to weeks.
Developed and deployed a comprehensive quantitative model to optimize the asset listings pipeline. Integrated internal trading data with external sources via APIs and custom web scrapers, pulling on-chain and off-chain metrics: social engagement, user account growth, trading volume, liquidity.
Python for modeling and feature engineering; SQL for querying and storage. Applied Random Forests and Gradient Boosting to rank candidate assets against historical performance and competitor benchmarks. The data-driven scoring system materially increased both the number and the quality of assets listed on Kraken.
05.7 kraken · strategy
Executive strategy.
Payward / Kraken · Head of Quant Research & Product Strategy
Built a team to apply innovative approaches in data structuring and analysis, using Python for quantitative modeling, ML, and statistical analysis to study user trading behavior, portfolio composition, and platform inefficiencies, optimizing product strategy and platform functionality to drive substantial revenue lift.
Partnered with cross-functional teams and leadership to define changes, design experiments, analyze results, and ship across products. Designed and automated reporting tools to surface complex data and ambiguous strategic problems for leadership, executives, and the board.
Built the data narrative and investor deck behind a successful, oversubscribed funding round. Owned narrative, modeling, metric selection, cohort analyses, growth drivers, and deck construction.
Building agentic infrastructure across the company. Workflows that span compliance, capital and risk, M&A, banking ops, and the cross-functional teams that depend on them. The connective tissue underneath: wiring up the internal systems that don’t talk to each other so the manual processing in between disappears.
The work is multi-agent system design. Workflow decomposition into orchestrated agent graphs with tool calling, structured output, RAG over internal data, and retrieval validation. MCPs as the standardized interface between models and internal systems. Eval harnesses on every agent path so regressions get caught before deploy. Human-in-the-loop where a decision needs accountability, autonomous everywhere else. Net effect: hours of manual processing per task removed, without losing oversight.
If anything here resonates, AI and agentic systems, automation, the operations side of finance and crypto, or just an interesting systems problem you’re thinking through, I’m always happy to connect.
