A SaaS habit that breaks in hardware.
SaaS teams ship a v1 and iterate against telemetry. Hardware teams can't — the BOM is frozen weeks before users see the device, and the next iteration is six months away.
Product leader · AI-native PM
For 13+ years, Ankit has led product across AI platforms, B2B SaaS, and hardware-software systems. He brings seasoned product leadership to anyone looking to build — whether you are a solo creator, a founder, or an established team navigating high ambiguity.
For anyone determined to take a concept to market, Ankit is the partner who drives execution. He excels at translating complex tech concepts into structured roadmaps, working prototypes, and products the market actually adopts.

What does Ankit do?
Ankit steps in to align your execution with real-world market realities. He helps you optimize spend, boost builder efficiency, and ensure every decision is directed toward products that actually sell.
Bringing senior leadership to the earliest stages, Ankit breaks the product down into shippable phases, delivering a structured roadmap built around real business constraints, not wishlists.
Usually, the specification is the problem, not the engineers. Ankit rebuilds the product brief, sharpens the core priorities, and removes the structural friction slowing down the ship date.
Having shipped core AI inside a real-world legal product (Cognizer), Ankit knows exactly how to separate true utility from hype. Whether you are embedding AI into a product or using it to automate complex workflows, he identifies the precise use cases that drive impact, and those that will simply burn your budget.
Operating directly within this product seam, Ankit bridges the two worlds. He translates the hardware lessons SaaS teams miss, and brings the platform decisions hardware teams skip.
You have the capital and the ambition to launch, but zero time to manage day-to-day product execution. Ankit steps in as your fractional product partner, protecting your time by driving the development forward, managing builders, and ensuring your vision gets built correctly without pulling you away from your primary career.
Selected work
Products and projects, mapped onto three circles. Most of what's interesting lives in the overlap. Scroll — the diagram tracks along — or click any region.
Launched Stack61 — a new low-five-figure ARR stream from existing clients — and re-platformed the field app, lifting mobile MAU ~50%.
petroit.com ↗Led discovery through launch — interviews, prototypes, signup experiments — then called the sunset when PMF didn't come, freeing the team.
Built for Proost Beer (Grano69 Beverages) — a Chrome extension that moves excise indents off the UP portal into ADB automatically; in daily operator use at v1.6.
bhangar.net/excise-connect ↗Took a dating platform from concept to launch — expectation-first matching with intuitive filters for an active community.
salgamosapp.com ↗Built and launched with family — landing, app, and WhatsApp flow live at aajkirasoi.com, planning real weeks of meals.
aajkirasoi.com ↗Took it from idea through ODM selection to small-batch production — early revenue, early customers, and a patent pending.
themagmirror.com ↗Led the pivot from general AI assistant to legal AI — ~98% F1 across 100+ labels and six-figure ARR per enterprise customer.
cognizer.ai ↗Built end to end, local-first — AI reads statements and receipts, splits every charge, and keeps an exact settlement to review in chat.
Built the menu-audit pipeline for restaurant operators — live menus scored, findings shipped as operator-reviewed action plans.
magna-carta.ai ↗Took modular pod storage from sketch to working hardware — E-ink label modules, an app, and a pre-launch funnel tracked daily.
podsque.com ↗Spec'd and shipped the household's control display — a zero-scroll, bilingual kiosk on a Pi, on the wall since July 2026.
Working on it.
Solved using agents
A voice-notes-to-Notion pipeline, Kickstarter research tools, a contact-enrichment engine, a device simulator — each a real problem handed to agents and shipped as a working repo. Clone the repo; run the solution.
github.com/bestankit-sudo→Each one a real problem, solved end to end. What you clone is what runs.
Plaud voice recordings become clean meeting notes — summaries, action items, named speakers — on your own Mac. Local Whisper transcription and voiceprint speaker ID; only the transcript text ever leaves.
Python · local Whisper (MLX) · Notion or local · clone & ./run
Given a list of Kickstarter campaigns, finds the best reachable contact for each company. AI validates every data point before it's written to Notion — precision over recall, credits spent only where they count.
TypeScript · Apollo · Brave Search · Notion
Tracks a Kickstarter pre-launch "Notify Me" follower count daily — the number Kickstarter's own dashboard won't chart — so ad spend can be read against real growth.
Apify actor · JSON / CSV output · scheduled daily
An interactive simulator for Podsque's hardware modules: e-ink display images at candidate panel resolutions, battery life across power budgets — decided before any hardware is produced.
Single-file HTML · no build step · live at podsquebatterysimulator.vercel.app
Given a list of companies, finds the best reachable contact for each — LinkedIn, verified email, or company channel — with AI validating every data point before a paid credit is spent. The generalized sibling of kickstarter-enrichment.
TypeScript · Apollo · Brave Search · Notion · MIT
Work with Ankit
Engagements ordered by intent.
Ankit is open to leadership roles — Director of Product, Head of Product, AI Product Leader, Platform Product Leader — where the work spans platforms, SaaS, and hardware.
Ankit comes in early to map products into shippable phases and anchor roadmaps to real constraints, not wishlists. The focus is on fixing the specification, which is usually the problem rather than the engineers.
Having shipped AI inside a real legal product, Ankit cuts through the market noise to identify which workflows AI will actually move the needle on, and which ones simply burn budget.
For those building something in the seams who want to compare notes with Ankit. A standing invitation to connect directly; no agenda, no deck.
Ankit mentors anyone going AI-native — product managers, founders, and builders. The topics are published free at mentorship.bhangar.net — no course, no cohort — with 1:1 mentorship for those who want the method applied to their own work.
Notes from the seam
Field notes from building across the seams — pinned as they land.
SaaS teams ship a v1 and iterate against telemetry. Hardware teams can't — the BOM is frozen weeks before users see the device, and the next iteration is six months away.
Platform PRDs describe capabilities; hardware PRDs describe parts. When a product spans both, the PRD that survives contact with the team is the one that lists both on the same page.
A horizontal AI tool wins demos and loses trials. The fix isn't a better model — it's a product surface that lets the model be tentative in public.