HumanAli. Human judgment. Artificial leverage.

I turn complex workflows into products people use.

10+ years building across AI, fintech and operational SaaS.
From discovery and prototypes to launch and go-to-market.

Get in touch Explore my work

More selected work

Across AI, automation and operational SaaS.

02UpEstateCRM-led B2B · Launched03CreatomaticAI listing videos · Public beta04Time DoctorBenchmarks AI05Process MetronomeNo-code ProcessOps06Lemon ClinicAI recommendations · Consultancy

02 / UpEstate · Launched · Spain

From internal real-estate tools to a CRM-led B2B business.

Concept to launch · Product & go-to-market

UpEstate property-publication interface showing listings, distribution channels and AI content suggestions
UpEstate · Property publication and distribution.

Origin

I turned the tools we built and lessons we learned in Madrid into a B2B product for real-estate operators in Spain.

Product Bet

Turn the workflows we developed while operating in Madrid into a repeatable product other real-estate operators could use. Package the CRM with training and a franchise model so technology and day-to-day operations work together.

Execution

A connected workflow for contacts, listings, sales pipelines and documents, with portal publishing, centralized messages, AI-assisted photography and automation, valuations, buyer matching and mortgage referrals. I led the project from concept to launch and created the go-to-market demos, courses and training videos.

Enabling operators

Acquire inventory through valuations, multi-portal publishing and integrated websites. Manage transactions through contacts, properties, pipelines and AI follow-up. Enable operators through demos, training courses and launch support.

Outcome

Built and launched the CRM-led offering, including onboarding, courses and training videos.

Visit UpEstate (opens in a new tab)

03 / Creatomatic · Public beta

One listing. A video ready to share.

Product Builder

Creatomatic video editor with property preview, audio controls and editable scenes
Creatomatic · Listing-video review and editing.

Problem

Every new property listing creates another content job. Agents need video, captions, and regular social posts, but manual editing competes with client work and generic content makes listings blend in.

Product Bet

Use the listing as the starting point. One property URL feeds a workflow that selects a pitch, language, and buyer profile, then generates a video and caption ready for review or publishing.

Execution

Built and launched the end-to-end beta: public listing ingestion, property and media extraction, video and caption generation, market and audience controls, and three delivery modes. Agents can run on autopilot, review before publishing, or export an MP4 and caption for any channel.

Launch Status

The multilingual beta is live. Agents can test it with their own inventory and receive a reviewable listing video from a public property URL.

Visit Creatomatic (opens in a new tab)

04 / Time Doctor

Make workforce data mean something.

Senior IC Product Manager, Experimental and Innovations team

Time Doctor total-hours percentile chart comparing teams against AI-matched profiles
Benchmarks AI · Peer comparison view

Problem

Leaders had internal productivity data but no credible peer context for judging whether team patterns were healthy, risky, or improving.

Product Bet

Use behavioral data to compare teams across industries, giving leaders a clearer view of how their patterns compare and where to investigate.

Execution

Led discovery and MVP delivery in under 3 months. We tested early prototypes on real customer data, then used pilot behavior and feedback to refine the experience and decide whether a larger build was worth funding.

Impact

Lifted engagement across pilot accounts by 75%. The pilot also gave us enough evidence to stop two weaker initiatives and redirect ~50% of the squad’s capacity.

Explore Benchmarks AI (opens in a new tab)

05 / Process Metronome

Give repeatable work a reliable rhythm.

Product Manager and Co-founder

Process Metronome operations dashboard showing process status, running instances and cycle-time charts
Process Metronome · Operations dashboard.

Problem

Operations teams relied on spreadsheets, tribal knowledge, and manager coordination to keep repeatable processes running across complex organizations.

Product Bet

Give operations teams a no-code way to map organizational assets, assign recurring work, handle exceptions, and keep day-to-day execution aligned with their standards.

Execution

Launched the first MVP in about a month, then added self-serve configuration, workflow templates, and workload planning based on direct feedback from customer operations teams.

Impact

Reduced node and asset setup time by 60%, enabled self-serve onboarding, and expanded the product across 3 enterprise clients. It became one of the company’s most frequently demonstrated capabilities.

Visit Process Metronome (opens in a new tab)

06 / Lemon Clinic · Consultancy

Start with the goal, not the catalogue.

Product Consultant

Lemon Clinic mobile experience showing personalized products and a guided Lemon AI recommendation conversation
Lemon Clinic · Consultancy AI recommendation experience

Problem

Users knew the outcome they wanted but struggled to translate broad beauty and health goals into relevant treatments, products, and next steps.

Product Bet

Use a guided AI conversation to clarify intent and match users to relevant catalogue options, giving the clinic a more useful starting point than navigation by treatment name.

Execution

Built the recommendation experience around goal selection, chat-based refinement, treatment and product matching, and a shopping-style interface for follow-through.

Impact

Increased recommendation engagement and reduced discovery friction by guiding users from broad beauty goals to relevant treatments and products through conversational AI.

Visit Lemon Clinic (opens in a new tab)

How I work

Clarity comes from getting close to the work.

Strategy, hands-on building and commercial ownership. I work through the whole system, from the first customer conversation to the decision about what ships next.

  1. Find the problem worth solving

    Start with the user decision or business bottleneck creating urgency. Cut ideas that won’t change behavior or remove a real constraint.

  2. Prototype in real context

    Use customer data, source material and realistic edge cases to test whether a prototype is useful before building the production version.

  3. Ship a working first version

    Build the smallest version users can complete a task with, then learn from how they use it.

  4. Make the investment decision

    Let behavior, operational results and customer feedback determine whether to deepen the bet, change it or stop.

  5. Automate what repeats

    Turn recurring manual choices into rules, tools and self-serve steps so growth doesn’t require the same growth in coordination.

Let’s work together

Let’s build something useful.

Get in touch