Consumer language-learning products treat every learner the same. Vocabulary arrives as generic word lists and standardized drills, disconnected from the memories, places, and experiences that make words stick. Retention suffers not because learners lack motivation, but because the material feels abstract — studied in isolation from lived context. For Abbett Labs, the business problem was sharper: the category rewards scale through standardization, yet the science of memory rewards personalization. Building a consumer product that could deliver one-to-one relevance without a tutor for every learner meant rethinking the product architecture from the ground up — not bolting personalization onto an existing curriculum engine.

I co-founded Abbett Labs and led product strategy to design a Human Learning System — a product architecture where personalization is the core mechanism, not a premium feature. The approach anchored vocabulary to personal photographs: learners capture words in the context of their own lives, and machine learning generates individualized lessons at scale. My role spanned product vision, learning experience design, and go-to-market narrative — translating cognitive science principles into a mobile product language investors and App Store reviewers could evaluate. I defined the learning loop (photo context → vocabulary capture → spaced retrieval → reflection), shaped the ML-assisted content pipeline, and positioned the product as evidence that scalable education does not require standardized education. The system made retention measurable: every lesson tied to personal context, every session generating signals the engine could adapt to.

Consumer education sits at a volatile intersection — cognitive science on one side, mobile behaviour and AI-assisted content generation on the other. Abbett Labs entered a market dominated by established players with massive content libraries and brand recognition. The team was small, the runway finite, and the hypothesis ambitious: that photo-anchored, ML-generated lessons could ship as a polished consumer experience, not a research prototype. Investor conversations and App Store launch required a product narrative that could survive scrutiny from both financial and pedagogical angles. The competitive landscape rewarded daily engagement metrics; the scientific literature rewarded context-dependent encoding and spaced practice. The strategic question was how to make those forces align in a single product architecture.

The product shipped on the Apple App Store, validating the core hypothesis: personalized vocabulary learning could move from research concept to usable consumer experience. Photo-anchored lessons replaced generic word lists. Machine learning generated individualized content without manual lesson creation per learner. KPIs tracked engagement against generic-list benchmarks, retention across spaced-retrieval intervals, and the ratio of learner-generated context to system-generated drills. The launch demonstrated that personalization at the architectural level — not the marketing level — could sustain a consumer product in a crowded category.

• Personalization without retention measurement is experimentation, not strategy — design the feedback loop before the feature roadmap.
• Learning systems should anchor content in lived experience; generic curricula optimize for content production, not memory formation.
• AI belongs in the learning engine that adapts delivery — not as language on the landing page that promises intelligence the product cannot demonstrate.
• Consumer education products must reconcile cognitive science with distribution economics; the architecture is where that reconciliation happens.

Grounded in context-dependent memory, spaced retrieval, and adaptive pathway design — principles from cognitive science applied to consumer product architecture. The engagement tested whether established findings on encoding specificity and distributed practice could survive the constraints of mobile UX, ML inference cost, and App Store discovery.