Insights · Notes from the field

Things we've learned shipping AI to production.

Essays on evals, compliance posture, postmortem culture, and the operational craft of zero-to-PMF engagements.

Featured · Compliance

PHI minimization in RAG pipelines.

Chunking, redaction, retrieval filters, and eval hooks so regulated AI answers without stuffing patient data into every context window.

Yash NerkarJul 30, 20269 min

Compliance

DPDP consent UX that founders can actually ship.

Practical notice, purpose limitation, and consent patterns for AI products selling to Indian regulated buyers — without freezing the roadmap.

Yash Nerkar9 min

Compliance

BAA before PHI: the sequencing most teams get wrong.

A decision tree for LLM vendors, subprocessors, and when synthetic data is enough — before anyone pastes real patient data into a prompt.

Yash Nerkar8 min

Compliance

HIPAA vs DPDP: one product, two postures.

Shared primitives for regulated AI — access control, logs, minimization — and the jurisdiction-specific deltas US and Indian buyers actually diligence.

Yash Nerkar10 min

Compliance

Choosing an LLM vendor under regulated constraints.

BAA availability, retention defaults, region options, and eval portability — how Brandlabs shortlists model providers for healthcare and regulated buyers.

Harsh Vitra8 min

Compliance

Audit logging for every inference call.

What to log for HIPAA-aware LLM systems, what never to log, how long to keep it, and who can open the trail.

Yash Nerkar9 min

Evals

Safety evals for clinical and health chatbots.

How to evaluate refusal, escalation, medical-advice boundaries, and hallucinated dosages — before a health assistant meets real users.

Harsh Vitra9 min

Evals

Building a prompt regression dataset that ages well.

Golden sets, adversarial sets, and ownership so your LLM suite still means something after six months of prompt thrash.

Harsh Vitra8 min

Evals

Three eval dimensions: functional, qualitative, safety.

The Brandlabs Discovery Sprint pattern for setting LLM eval targets in weeks 1–3 — before feature thrash eats the timeline.

Harsh Vitra8 min

Evals

CI that blocks deploys on failing evals.

Sampling, cost caps, flake handling, and which LLM eval suites must stay required vs nightly — so production cannot silently regress.

Harsh Vitra8 min

Delivery

Who sits in an embedded AI pod.

FDE, senior engineers, product design, and compliance-as-needed — roles and anti-patterns for shipping regulated AI in 12–20 weeks.

Yash Nerkar7 min

Sector

How US healthcare startups should buy AI build partners.

Diligence questions on BAA fluency, eval discipline, and clinical workflow empathy — before you hire a studio to ship your AI product.

Yash Nerkar8 min

Delivery

The two-week Discovery Sprint, unbundled.

Inputs, outputs, kill criteria, and how Brandlabs maps a Discovery Sprint into a 12–20 week zero-to-PMF engagement.

Harsh Vitra7 min

Sector

Regulated-industry AI buyer’s guide (India).

Questions CFOs and compliance officers should ask before signing an AI vendor — DPDP, updates, liability, and evidence — not just a demo.

Yash Nerkar8 min

Delivery

Zero to PMF in 12–20 weeks: the phase model.

Foundation → build to alpha → production-harden → PMF hunt — with eval and compliance gates at each phase.

Harsh Vitra9 min

Sector

Trust UX for telehealth and women’s health platforms.

Privacy, language, and stigma-aware product patterns from shipping telehealth for women’s health — so care feels safe enough to start.

Yash Nerkar9 min

Sector

Safety engines for WhatsApp health assistants.

Escalation, localization, and content safety patterns for healthcare chat on WhatsApp — from shipping Myna Bolo in production.

Harsh Vitra9 min

Sector

Voice AI for low-literacy health access.

Latency, vernacular STT/TTS, and when voice beats chat — lessons from rebuilding Myna Voice AI for rural women’s health support.

Yash Nerkar8 min

Sector

AI for Indian MSME compliance (lessons from Correct).

Productizing 69,000+ obligations and regulations that change ~15 times a day — without turning an AI assistant into unauthorized legal advice.

Harsh Vitra10 min

Evals

Evals are the product.

Why we treat the eval suite as a first-class deliverable — and how it changes how we negotiate scope.

Harsh Vitra6 min

Delivery

Shipping on Fridays without breaking trust.

A pattern for weekly demos that are evidence, not theatre. What we ship, what we don't, and what we say out loud.

Yash Nerkar8 min

Compliance

Compliance as architecture, not paperwork.

HIPAA and DPDP shouldn't be a last-mile checklist. Here's how we bake them into the first commit.

Yash Nerkar10 min

Evals

Every postmortem becomes a regression test.

The cultural and technical loop that turns production incidents into permanent guardrails.

Harsh Vitra7 min

Delivery

What a forward-deployed engineer actually does.

Not a project manager, not a consultant. A senior engineer who lives in your Slack and writes the hardest code.

Harsh Vitra5 min