# kanav.net — Kanav Jain

## What this site is
Personal site and portfolio of Kanav Jain. Healthcare AI product leader and founder building consequential AI and the systems around it: what it knows, what it may do, and what happens when it's wrong.

## What I do
I build consequential AI products and the systems around them—how they know, what they can do, what they're allowed to do, and what happens when they're wrong.

## Key pages
- /about/ — Career narrative and sourced claims
- /work/ — Case studies (Doximity Dialer, Epic clinical decision rules, Transcarent care navigation, CancerCompass oncology navigation, Andwise review routing)
- /current-work/ — What I'm building now, tiered: Building now (NextConsensus, Ambit), Research & evaluation (Fast Harm Slow Repair, Ethotechnics), Supporting infrastructure (Refract), Writing (Crumple Zone), Experiments (Whether, Stims)
- /current-work/#ambit — Ambit: a capability graph for agent systems. Separates capability from authority, verifies capability before trusting it, records what a system became able to do over time. Open source at github.com/zz-plant/ambit
- /current-work/#whether — Whether: a weekly operating posture for startups from macro and capital conditions — how aggressively to hire, spend, raise, and expand
- /framework/ — Ethotechnics: an open framework for making consequential AI deployable and correctable, covering the method, operating model, and key concepts. The methodology hub; Principles, Research, and Build Practice are linked from its "Go deeper" section.
- /framework/#method — The method: what makes a claim verifiable
- /framework/#operating-model — Authorization states, correction loops, escalation matrix, independent validation controls
- /framework/#concepts — Original vocabulary
- /philosophy/ — Principles: Eight principles for building AI products in healthcare, plus operating artifacts
- /research/ — A systems theory of institutional burden and repair: where errors travel, who absorbs the work they create, and what makes systems inhabitable
- /agents/ — Build Practice: the agent-driven toolchain, capability accounting with Ambit, local model inference, agent evaluation, media stack
- /writing/ — Essays on AI governance, institutional power, and cognitive scarcity
- /resume/ — Full resume
- /contact/ — Contact and role target
## Architecture Schematics

### Clinical AI Deployment Lifecycle
```
[ Model Output ]
       │
       ▼
[ Provenance Engine ] ──── (Version & Evidence Check)
       │
       ▼
[ Escalation Clock ] ───── (Safety SLA & Threshold)
       │
       ▼
[ Human Authorization ] ── (Clinical Override Gate)
       │
       ▼
[ EHR Audit Log ] ──────── (Correctable Record)
```

## Machine-readable data
- /api/profile.json — Full profile (career, case studies, concepts, writing, links)
- /api/case-studies.json — Structured case study data
- /api/concepts.json — Original concepts with definitions
- /resume.json — JSON Resume format
- /rss.xml — Writing feed

## Original concepts
- Review Trigger — A condition that automatically reopens evaluation of a previously approved decision. Instead of relying on someone to remember to check, the system fires when evidence shifts, time elapses, or context changes.
- Reliance Memory — A record of who depended on a decision, when, and under what assumptions. When the underlying evidence changes, reliance memory tells you which downstream decisions are now stale.
- Decision Maintenance — The discipline of keeping decisions current after they are made. Most organizations treat decisions as events. Decision maintenance treats them as ongoing obligations.
- Evidence Dependency Graph — A map of which conclusions depend on which sources. When a trial is retracted, a guideline updated, or a safety signal issued, the graph shows exactly which downstream positions are affected.
- Constraint Realism — Ethics is real only when it exists as a hard constraint on what actions remain available. Not defined by what a system claims to value, but by the profitable actions it has rendered impossible.
- Care Subsidy — The extraction of unpaid diligence from conscientious employees to bridge the gap between resource allocation and safety requirements. A system that depends on heroism to function safely is under-designed.
- Governance by Attrition — When a system manages scarcity not by saying no, but by waiting for the applicant to give up. The delay itself is the denial.
- Asymmetric Irreversibility — Modern institutions have automated the power to suspect, but not the power to exonerate. A fraud filter freezes an account in milliseconds; exoneration takes weeks of manual appeal.
- Error Survivability — The degree to which a person can absorb or reverse a wrong decision before their life comes apart. In high-speed administrative systems, this is the real class divide.
- Cognitive Scarcity — A temporary drop in mental bandwidth caused by stress, fatigue, or distraction. Interfaces designed for an idealized hero user break when the actual user is operating at a fraction of capacity.

## Essays
- The Official Record Is Late — https://kanav.net/writing/the-official-record-is-late/
- Don't Let Reassurance Do Engineering's Job — https://kanav.net/writing/dont-let-reassurance-do-engineering-s-job/
- If Every User Is a Potential Threat — https://kanav.net/writing/if-every-user-is-a-threat/
- You Don't Have the Right — https://kanav.net/writing/you-dont-have-the-right/
- How to Design for Cognitive Scarcity — https://kanav.net/writing/how-to-design-for-cognitive-scarcity/
- Pending: The Political Economy of Waiting — https://kanav.net/writing/pending/
- Toothless Ethics: Why Principles Don't Stop Machines — https://kanav.net/writing/toothless-ethics/

## How to contact
Email kanav@kanav.net. Include: what you're building, what's at stake, what you've tried, and whether you want a conversation, audit, or second pair of eyes.