Available for new opportunities

Rabina Karki

I build as a

I take products from blank doc to market. Fast, lean, backed by data. AI-native builder with 5+ years shipping things people actually use.

Rabina Karki
AI Product Builder
1,000+ Signups
$0 CAC
5+
Years in Product
18%
Retention Lift
scroll
Product Manager AI-Native Products 0 to 1 Builder LLM Integration B2B SaaS Agentic Workflows Prompt Engineering GTM Strategy Product Manager AI-Native Products 0 to 1 Builder LLM Integration B2B SaaS Agentic Workflows Prompt Engineering GTM Strategy
About Me

The person behind
the product.

I don't build features. I build outcomes. My job is to figure out what matters, say no to everything else, and ship the thing that moves the number.

I co-founded OnePetz and took it from zero to Play Store in 6 months. No budget, no codebase, pure prioritization. Before that, I shipped 3 products at ConsultEdge.Global that drove the highest quarterly revenue growth in company history.

I live at the intersection of AI and product. Integrating LLMs into real user workflows, not just demos. Gemini, OpenAI, LangChain, Pinecone, all shipped in production.

LLM IntegrationAgentic Workflows0-to-1 BuilderAsync-firstUS Remote ReadyB2B SaaS
0
Years Building
0
Pre-launch Signups
0
Revenue Growth, One Quarter
Current Role
Co-Founder, OnePetz
AI Pet Super App, Head of Product
Location
India, Remote Ready
Open to US remote roles
Specialization
AI-Native Products
B2B SaaS, B2C, 0-to-1 Launches
Email
rabinakarki1497@gmail.com
Best way to reach me
Capabilities

What I work with.

🎯Product
0-to-1 LaunchesProduct StrategyRoadmappingPRD WritingGTMOKRsRICE PrioritizationKano ModelDiscovery SprintsUser Research
🤖AI and Tech
LLM IntegrationPrompt EngineeringRAGAgentic WorkflowsAI EvaluationOpenAI APIGemini APILangChainPineconeClaude
📈Growth and Analytics
A/B TestingFunnel OptimizationCohort AnalysisRetention AnalysisMixpanelGoogle Analytics 4SQLPython
🛠Delivery and Tools
FigmaAgileScrumJiraNotionStakeholder ManagementAsync CommunicationCross-functional Leadership
Case Studies

Things I've shipped.

Real products. Real decisions. Full breakdown of how I think and execute.

Consumer SaaSLLMPrompt Eng.Monetization16 Weeks

Canva Magic Design. Text to Design in 3 Seconds.

Solved blank-canvas paralysis for 200M+ non-designer users. Reduced creation time from 30+ minutes to under 5 seconds via hybrid AI stack, driving 18% retention lift and $30 to $60M incremental ARR.

18%
30-day retention lift
131%
Designs per user increase
$60M
Incremental ARR (est.)
Core Problem
  • 67% of users said: I wish someone could just make this for me
  • Users spending over 20 minutes on first design had 2x churn risk
  • 18% of support tickets were: How do I start a design?
  • $40M ARR at risk from plateauing core-user engagement
  • Competitive threat: Adobe Express launched free tool in 2023
AI Inference Pipeline
User Input "birthday poster" GPT-4 Prompt Expand vague input 1 to 2 sec Stable Diffusion Image generation 2 to 3 sec Post-Process Image to Layers 1 to 2 sec Editable Design Ready to use ~5 sec total Cache hit: top 100 templates load instantly, under 100ms
Baseline Testing
100 real briefs. DALL-E 3 vs SD. 70% acceptable, proceed hybrid.
Prompt Engineering
GPT-4 translates vague input to visual spec. Quality 70% to 89%.
Employee Beta
200 employees. Found mobile lag, cost explosion, moderation gap. Fixed.
Phased Rollout
1% to 5% to 25% to 100%. Caught non-English quality drop at 25%.
Key Tradeoffs
DecisionChosenWhy
ModelSD 80% + DALL-E 20%$4.4M vs $500M per year
Latency3 to 5 sec balancedFast and ok beats slow and perfect
MobileAsync rendering3x perceived speed
Paywall5 free per month8% conversion, 3x industry avg
Results
  • Day 30 retention 28% to 33% (+18%)
  • Designs per user 3.2 to 7.4 (+131%)
  • Premium conversion 12% to 19%
  • Cost per generation: $0.0089 vs $0.04 worst case
  • Pro designers used it for rapid ideation, a surprise finding
EdTechMultimodal AIVision + LLMQuality-Critical20 Weeks

Physics Wallah. Instant Doubt Resolution at Scale.

Replaced the tutor bottleneck with a multimodal AI. Students photo their problem; AI answers in under 1 minute with 88%+ accuracy. Eliminated a 40K-doubt daily backlog. Result: $17.5M ARR at 97% margin.

44x
Faster (45 min to 1 min)
$17.5M
New ARR from premium
97%
Margin vs tutor model
Core Problem
  • 50,000 doubts per day. 1,000 tutors can answer 10K/day max
  • 40K unanswered daily. 35% of those students churn in 30 days
  • Tutor cost $0.50 per doubt = $18M per year to close the gap
  • Crowdsourcing rejected: 20% wrong answers means brand damage
  • Photo to structured explanation is textbook multimodal AI
System Architecture
Photo Upload Student snaps problem Google Vision OCR + extraction Problem identified Subject Router Physics/Math/Chem GPT-4-Turbo Generate explanation 75%+ confidence Send instantly Below 75% Tutor review (8%) Student Under 1 minute
Baseline Testing
5K historical doubts. 88% accuracy at $0.020 per doubt. Proceed.
Subject Tuning
Chemistry 72% to 81%. Subject-specific prompts. Organic reactions special-cased.
Tutor Validation
20 tutors x 250 explanations. Physics 91%, Math 93%. Overall 88.2%.
Beta + Rollout
5K students to 1% to 100%. Blurry photo fix, spam filter, confidence tuning.
Unit Economics
Cost ItemPer Doubt
GPT-4 API$0.008
Google Vision$0.006
Tutor review (8%)$0.004
Infrastructure$0.003
Total AI cost$0.021
Previous tutor cost$0.50
Savings95.8%
Surprising Findings
  • Rural students adopted fastest: 45% vs urban 28%. AI became their only tutor.
  • 2 AM peak equals 9 PM peak. Night-before-exam cramming is real.
  • Harder questions to AI: 6.2 out of 10 vs 4.1 with tutors. No social judgment.
  • Confidence threshold at 75%: 80% over-flags, 70% under-catches.
RetailPersonalizationKnowledge BaseVet Partnerships16 Weeks

Chewy Pet Advisor. From Transaction to Relationship.

Moved Chewy from "Amazon for pets" to trusted pet health partner via AI chat that learns each pet's profile. Drove +32% AOV, +26% retention, and $340M incremental annual revenue at 95% margin.

+32%
Average order value
$340M
Annual incremental revenue
+26%
12-month retention
Core Problem
  • 62% of pet owners: I don't know if I'm buying the right products
  • 1.2 categories per customer. Food is 80%, 60% of profitable categories missed
  • Amazon launched same-day pet delivery at lower prices, a commodity trap
  • LTV gap: commodity buyer $400 vs engaged customer $1,800
  • Strategic need: discount supplier to trusted pet care advisor
Recommendation Pipeline
Customer Chat "7yr Lab, joint pain" GPT-4 + KB 50K products, 200 breeds drug interactions Safety Checks Breed-specific risks Top 3 Products Ranked + explained Cart + Checkout $52 to $68 AOV 73% add to cart Feedback loop: retrain knowledge base monthly
Problem + Data
50 interviews. 73% want AI advisor. KB architecture designed.
LLM + KB Build
GPT-4 + KB: 78% to 91% accuracy. 50K products, 200 breeds, drug database.
Chat Beta
AOV +28%, NPS 4.2/5. Fixed chat brevity, mobile UX, 20 to 3 product cards.
Full Rollout
1% to 100%. Tutorial added at 10%. Metrics stable at 30M users.
Revenue Math
MetricValue
AOV lift per transaction+$16 (+32%)
Repeat frequency+40%
Commerce incremental ARR$325M
Premium subscriptions (2M)$15M ARR
Total incremental ARR$340M
Annual AI cost$17M
Net margin95%
What Made It Defensible
  • Proprietary KB: 50 vet partners, drug interactions. Amazon can't copy in 6 months.
  • Vet referrals became inbound. 12% of new customers cited vet referral.
  • Multi-pet from day 1. 67% of customers have multiple pets. Single-pet = incomplete.
  • Emotional moat. Trust beats price competition every time.
Experience

Where I've been.

Co-Founder and Head of Product
OnePetz, AI Pet Super App
Nov 2024 to Present
  • Took OnePetz from blank doc to Play Store in under 6 months. All PRDs, full lifecycle, no prior codebase or budget.
  • Shipped AI pet health advisory on Gemini API. 4.2 out of 5 satisfaction on MVP, vet consultation completion up 18%.
  • Scaled to 1,000+ pre-launch signups at zero acquisition cost via organic community partnerships.
  • Declined 60%+ of Q1 feature requests to protect retention-driving core flows.
Fractional Product Manager
Independent Consulting
Mar 2024 to Nov 2024
  • Advised 3+ early-stage startups on roadmap, PRD development, and GTM strategy across B2C and SaaS.
  • Helped founding teams cut scope by 30 to 40% via RICE and Kano prioritization workshops.
  • Built user research pipelines translating interviews into structured backlogs with defined success metrics.
Technical Product Analyst
ConsultEdge.Global, B2B SaaS and eCommerce
Oct 2022 to Mar 2024
  • Launched 3 products on XCommerce, driving 20% revenue growth in Q1 2023. Highest single-quarter in company history.
  • Increased XeDigital.ai activation by 15% via A/B testing. Satisfaction scores up 25%.
  • Achieved 15% feature adoption growth QoQ with 100% on-time delivery.
  • Mentored 2 analysts, both promoted to product roles within 6 months.
Product Analyst
Legacy IoT, Global IoT Platform
Feb 2020 to Aug 2021
  • Promoted from Support to Product Analyst in 6 months, earning direct PM ownership under the CTO.
  • Ran 100+ user interviews. Roadmap drove 15% platform adoption growth in 3 months.
  • Rebuilt onboarding: engagement up 30%, time-to-adopt down 15%.
  • Reduced critical defect rate by 20% via structured QA feedback loop.
Testimonials

What people say.

"

Rabina is one of those rare people who makes a team better just by being in the room. At ConsultEdge she always knew what to prioritize, when to push back, and how to get everyone moving in the same direction. She didn't wait to be told what to do. She figured it out.

HA
Himanshu Ahuja
Manager, MNP Spark — Direct TL at ConsultEdge Global
"

Building OnePetz with Rabina was one of the best decisions I made. She owned the product end to end, asked hard questions from day one, and never let us ship something we weren't proud of. Her clarity under pressure is something I haven't seen in many people.

CS
Charan Surya
Product Manager, Co-Founder at OnePetz
"

Rabina thinks like an engineer but communicates like a product person. Our brainstorming sessions are always sharp and grounded. She challenges assumptions without being dismissive and always comes with a point of view, not just questions.

SS
Shweta Singh
Senior Software Engineer, Accenture
"

I saw Rabina grow from someone learning the ropes to someone who could run a full product cycle independently. What impressed me most was her user empathy. She never built for the sake of building. Every decision tied back to a real problem.

PK
Pankaj Kumar
Founder, Legacy IoT — Early Mentor
"

Working alongside Rabina at ConsultEdge gave me a front-row seat to how she operates. She is methodical, she listens well, and she has a gift for simplifying complex problems into something the team can actually execute on.

SB
Shweta Bisht
Senior, ConsultEdge Global
"

Rabina helped me think about my projects in a completely different way. She asked questions I hadn't thought to ask, helped me define who I was actually serving, and gave me a roadmap I could actually follow.

SK
Sarvesh Kushwaha
Senior SDE 3, Fluor
Value Add

What you get.

🚀
Ship Fast

Zero to launch in weeks. I cut scope ruthlessly and move before conditions are perfect.

🧠
Think in Systems

I see how decisions cascade. Roadmap, GTM, structure. I connect dots others miss.

📊
Data Before Opinion

Every call grounded in evidence. I build feedback loops so data stays fresh.

🤝
Lead Without Authority

Engineers, designers, execs, aligned without a title. Clarity and trust are my tools.

Contact

Let's build something.

Open to full-time PM roles, fractional engagements, and advising early-stage teams. If you are building something ambitious, I would love to hear about it.