The portfolio recruiters can't ignore.
Paste a job description. Drop your LinkedIn. In 96 seconds, walk away with a brand-matched portfolio site + a PM-style pitch tailored to that exact role.
Research agent
Reading the JD
Brand agent
Pulling brand colors + fonts
Narrative agent
Writing your story
Pitch agent
Drafting your PM-style pitch
Code agent
Generating Next.js + zipping
5 agents · 1 portfolio
Each agent does one job. Together they ship a portfolio in 96 seconds.
Research agent
Scrapes the JD, parses career level, infers the right pitch stance.
Brand agent
Pulls real brand colors + fonts via Brandfetch. Your site looks native to the company.
Narrative agent
Writes your headline, About, cover letter, and role-specific resume bullets. Zero generic fluff.
Pitch agent
Optional PM-style pitch — problem, hypothesis, solution, metrics. Builder · Analyst · Customer · Strategist.
Code agent
Generates a buildable Next.js project + zips it for self-deploy. Done in milliseconds.
See it for real
An actual generated site — not a mockup.
A real JobMagnet run: a Group Product Manager application to Duolingo, themed in Duolingo's own brand. Scroll inside the frame, or open the full site.
Akshey Walia
Product and AI leader who grew Glance across 8 markets and 250M+ pre-installed devices — interested in making Duolingo a daily habit for millions of learners in India.
Why I'm a fit
Built for this role, not just any role.
100K to 6M devices
At InMobi, I led Glance's US 0-to-1 growth from no playbook to Verizon and Motorola distribution, combining research, GTM, compliance, and product integration. That maps directly to defining Duolingo's India strategy while adapting global levers for local learners.
5-PM team
I built and led a 5-PM team from zero, set ownership across 8 markets, and promoted 2 PMs to Senior PM in 18 months. Duolingo needs a GPM who can grow a team while keeping urgency and quality high.
3X activation
I use funnel data, user research, and design iteration to improve habit-forming products, including activation, reactivation, retention, and engagement experiments on the lock screen. That is the same muscle needed to make learning sticky in India.
A pitch · builder stance
What I'd ship at Duolingo
The problem
Duolingo's homepage promise is broad: quick lessons that build real-world communication skills across reading, writing, listening, and speaking, while the English course itself now includes job-interview content buried inside later B1/B2 sections. For an India growth user with an urgent English-speaking job outcome, the observable course-entry flow lacks a focused way to declare that mission and start a speaking-heavy habit immediately.
Hypothesis
If we add a lightweight "Job interview speaking" mission at English onboarding and the course section selector, then D7/D30 retention and lesson starts improve because high-intent learners get a concrete daily loop tied to a career moment instead of searching through the general path.
Proposed solution
Ship a next-sprint experiment on the English course start screen: after language selection, add a "Choose your mission" step with the default path plus a "Job interview speaking" card for India English learners. Selecting it creates a 14-day mission lane on the learning path with 8-minute daily sessions: listen to an interview prompt, tap-to-speak an answer, get fluency/pronunciation feedback, then earn normal XP/streak credit; reading and writing exercises remain available but are de-emphasized. Reuse existing B1/B2 interview scenarios and monolingual English-mode content, but package them as a temporary overlay that can hand the learner back to the main path after the mission.
Metrics to track
Tradeoffs
- A mission lane may fragment the simplicity of Duolingo's single learning path.
- Speaking-first sessions can be noisy in public or low-bandwidth contexts common for mobile learners.
- Career framing may raise expectations beyond what a short course can deliver.
- Repackaging existing content is faster than net-new curriculum, but may feel less personalized without scenario selection.
Guardrails
- Keep the mission optional and reversible; never block the standard English path.
- Cap sessions at 8 minutes and preserve streak/XP mechanics so it strengthens the daily habit loop.
- Localize prompts for Indian job contexts without stereotyping accents or penalizing valid Indian English pronunciation.
- Require clear learner consent before microphone use and provide a listen-only fallback.
- Measure learning-path progression cannibalization so mission gains do not come from lower overall progress.
Evidence
Before: one generic English path; proposed: an optional job-interview speaking mission at course start.
Selected work
Where I've worked
- InMobi·Group Product ManagerJanuary 2023 - April 2026
- Led Glance AI lock screen GTM and 0-to-1 workstreams across 8 markets on a platform pre-installed on 250M+ devices.
- Grew US distribution from 100K to 6M devices in 18 months through research, Verizon and Motorola integration, and CCPA-aware product strategy.
- Improved activation from 25% to 75% by redesigning onboarding with Lottie animations, sharper value props, and funnel-level segmentation.
- Raised reactivation from 5% to 25% using weather-triggered nudges for dormant users while preserving uninstall choice despite a 20% funnel drop.
- Built a 5-PM team from zero, promoted 2 PMs to Senior PM, and set ownership across 8 markets.
- InMobi·Senior Product ManagerJanuary 2022 - December 2022
- Drove a lock-screen UI experimentation SDK with engineering, enabling A/B testing inside OEM SystemUI APK and lifting engagement 20%.
- Led Motorola's first Glance SystemUI integration, creating reusable patterns later used for Samsung, Sharp, Xiaomi, and telco partnerships.
- Scaled Realme from 2M to 25M DAU by solving Android GO constraints, OTA rollout, and new-device out-of-box integration.
- Launched Brazil from zero local presence, adapting GTM, Portuguese localization, LGPD compliance, and carrier billing to reach 10M devices.
- Magicbricks (Times Internet)·Product ManagerAugust 2015 - December 2021
- Built Pay Rent from zero to INR 50Cr GMV with Razorpay integration, 3 gateway routing, bank cashback wiring, and settlement improvements.
- Raised payment success from 95% to 99% by onboarding 2 additional gateways and routing transactions during failure spikes.
- Validated 5 property-service verticals manually before automation, scaling from 10 to 500 orders across 10 partners.
- Increased home-loan lead generation 40% by redesigning capture flows and onboarding HDFC, ICICI, Standard Chartered, IDFC, and Kotak.
- Ocus Group·Associate Product ManagerApril 2012 - March 2015
- Led 2 product launches into new geographies, generating INR 5Cr in pre-launch revenue.
- BPTP Limited·Assistant ManagerNovember 2010 - March 2012
- Drove INR 100Cr revenue through key account management and improved onboarding efficiency by 30%.
- Newbridge·Assistant ManagerJuly 2008 - October 2010
- Launched a new revenue channel and improved lead generation by 80% through process redesign.
About
I'm a product leader who has spent the last four years growing Glance across markets, partners, and user behaviors that were still being understood. I like roles where strategy has to become shipped product: research, local insight, crisp specs, launch, analysis, and iteration. At Duolingo, I'd bring experience building PM teams, adapting global products for local markets, and using data plus design judgment to make daily engagement feel natural instead of forced.
median end-to-end generation
brand-matched to the company, not a stock theme
self-deploy to Vercel · Netlify · anywhere