Skip to content
Ajay Krishnan.

Engineer · Lawyer · Builder

Two decades shipping software. A licensed lawyer the whole time.

I build production AI systems end to end — and I own the governance, privacy and IP surface most engineers route around.

San Francisco · shipping software since 2004 · 22 years

The short version

Most people have one of these. I have spent years doing all three at once.

Most executives stopped coding years ago. Most engineers never lead at the C-suite. Almost none are members of the bar.

Ship

Production code, daily, across TypeScript, Python and Ruby. Multi-agent orchestration, tool-calling, RAG, and the evals that keep them honest.

Lead

Engineering organisations of 21, 18 and 11. Three startups, two exits. Twice a General Counsel, twice a CTO — usually at the same time.

Defend

Called to the Ontario bar in 2012. When a General Counsel asks whether the thing can be trusted, I can answer as a lawyer and as the person who built it.

Receipts

The numbers behind that.

Restaurants
27,000 Restaurants Burger King, Tim Hortons and Popeyes, where I led legal, privacy and PCI for the first mobile order-and-pay apps
Countries
100+ Countries The footprint those brands operate across
Engineers led
21 Engineers led At City Lifestyle, through three team leads, while also serving as General Counsel
Built to acquisition
0 → 18 Built to acquisition Lobo Genetics, from co-founding through to the exit
Agent-authored PRs
92% Agent-authored PRs At AI Growth Labs, where the engineering organisation itself is the thing being rebuilt
Series A quarterbacked
$12M Series A quarterbacked As General Counsel and Corporate Secretary at Raven.ai

The long version

Twenty-two years, in the order they happened.

  1. 2004 — 08

    Co-Founder at Savvica

    Toronto

    An e-learning platform, and the start of shipping software for a living.

    • Founder
    • E-learning
  2. 2006 — 08

    Tech Lead at Infusion, later Avanade

    Toronto · London · Dubai

    Hired and led a net-new team of 11 engineers at Barclays Capital on a $12M budget. Launched the Dubai office. Built the hotel and restaurant reservation system behind Jumeirah.com.

    • Team of 11
    • Capital markets
  3. 2010 — 12

    Law Articles at Davies Ward Phillips & Vineberg

    Toronto

    Junior counsel on a 5–4 win at the Supreme Court of Canada. Called to the Ontario bar in 2012.

    • Supreme Court of Canada
    • Called 2012
  4. 2016 — 18

    Counsel, Operations & Technology at Restaurant Brands International

    Toronto

    In-house at Fortune 500 scale for Burger King, Tim Hortons and Popeyes — 27,000 restaurants across 100+ countries. Negotiated technology, commercial and supply-chain contracts, and led legal, privacy and PCI compliance for the brands' first mobile order-and-pay apps.

    • 27
    • 000 restaurants
    • PCI
    • Privacy
  5. 2018 — 21

    Co-Founder, CTO & General Counsel at Lobo Genetics

    Toronto

    Built the full stack from zero to acquisition — a consumer genetic testing platform and a two-sided marketplace — and scaled the team 0 to 18. Stood up the legal function from scratch, owning IP and venture financing, and steered the product through Health Canada approval.

    • 0 → 18
    • Exit
    • Health Canada
  6. 2021 — 24

    Technology & Commercial Counsel at Caravel Law

    Toronto · fractional, concurrent with operating roles

    Counsel to technology companies from seed to enterprise on commercial contracting, SaaS and IP licensing, open-source compliance, privacy under PIPEDA and GDPR, and AI governance. Quarterbacked a $12M Series A as General Counsel and Corporate Secretary at Raven.ai.

    • AI governance
    • $12M Series A
  7. 2022 — 23

    VP, Trade Technology at Fishtail

    Toronto

    Architected a fintech trade-finance platform — Rails, PostgreSQL, Hotwire and GKE, with a real-time document-processing pipeline — and built the international expansion playbook across the US, Mexico and Colombia.

    • Fintech
    • 3 markets
  8. 2024 — 25

    CTO & General Counsel at City Lifestyle

    Kansas City

    Led a 21-person engineering organisation through three team leads across a 300-magazine network, and owned the entire legal function alongside it. Designed and ran the company's AI transformation, driving adoption across 220 franchises through demos, training and enablement — shipping tools to people who had not asked for them, then iterating until they were daily habit.

    • 21 engineers
    • 220 franchises
    • AI adoption
  9. 2025 — now Now

    Founding Engineer at AI Growth Labs

    San Francisco

    Building TravelAgent.ai end to end across TypeScript, Python and Ruby — multi-agent orchestration, tool-calling across frontier models, and RAG over a Neo4j knowledge graph, with evals gating every change. I spend as much time on how we measure what the agents produce, and on finding where they quietly fail, as on the build itself.

    • Agents
    • Evals
    • Neo4j

Three educations

I kept going back.

01

Engineering

Full-stack, and still writing it. TypeScript, Python, Ruby. Rails and Next.js. Agentic systems in production.

BASc, Engineering Science — University of Toronto

02

Law

Called to the Ontario bar in 2012. Corporate, commercial and technology practice: licensing, privacy, IP, AI governance.

JD — University of British Columbia

03

Business

Two exits, three startups, and the go-to-market side of a platform rollout across three national markets.

MBA — DeGroote School of Business, McMaster University

Things you can open

I would rather show you than tell you.

  • 01

    Legal AI · built alone

    Blackline Legal

    Reads a Canadian employee's situation against the statutes and the case law, tells them what they are owed, and drafts the letter they can hand to HR. Trained on a decade of my own employment-law practice. I built the privacy architecture myself: each case stays isolated, and nothing anyone enters trains the model.

    • LLM-native
    • Employment law
    • Per-user isolation
    Open it

    blackline.legal

  • 02

    AI news

    New Grok Times

    A daily newspaper written entirely by AI across twelve bureaus. Every claim cited, every source linked, readable offline.

    • Twelve bureaus
    • Source-cited
    • PWA
    Open it

    ngtimes.org

  • 03

    AI policy

    Better Canada

    Every federal Canadian regulation reviewed through a free-market lens — 3,816 of them, each given a verdict. Total infrastructure cost, zero.

    • 3
    • 816 regulations
    • Free-tier LLMs
    Open it

    bettercanada.krishnan.ca

  • 04

    Audio platform

    Rajam

    Voice-first recording that treats a voice like it matters. Offline support, real-time waveforms, one-link sharing.

    • PWA
    • Offline
    • Hotwire
    Open it

    rajam.krishnan.ca

  • 05

    Experiments

    Edge AI Agents

    Small language models running autonomously on ESP32 and Raspberry Pi Pico. Local inference, no cloud, physical-world interaction.

    • ESP32
    • Pico
    • MicroPython
    In progress

In my own words

What I actually think about all this.

“I have written the job description, sat in the loop, made the call, built someone's first week with no playbook, and had the promotion conversation that went badly. When I say I have felt the cost of bad people practices, it is from the inside — some of them were mine, and I never had the right tools.”
On building for people
“Legal work is shifting from drafting documents to directing and supervising agents, and the lawyers who win will be the ones who can build and govern those workflows.”
On where this goes
“At Lobo we took a consumer genomics product through Health Canada; before that I owned privacy and PCI for payment apps at scale. In those rooms, careful deployment was part of the engineering, not a tax on it.”
On regulated work
“I own it end to end, from the model to the product, and what I care about most is whether it actually works for the person on the other end, not what demos well.”
On what matters

Say hello

I read everything. I reply to most of it.

Best over email, or put something in the calendar and we can talk properly.

In the agent era, range is leverage.