Andressa K.

Grounded in research, delivered in code. I own the product from question to merge.

Is Andressa a good fit for startups?

Selected cases

Industry Talks

Presented our new product, Predict.PM, at Manifest — Oakland, CA

2025

As one of its creators, I shared the stage with our CEO and Futuur's market maker to introduce it to a crowd of superforecasters and AI enthusiasts sharpening their forecasting skills.

Presented our new product, Predict.PM, at Manifest — Oakland, CAPredict.PM product screenshot

What I'm Building

Marketing content assistant

2026

A Slack-triggered bot that turns a list of topics into publish-ready social posts: it researches and summarizes the news, correlates it with Futuur's prediction markets via LLM, then generates on-brand images and copy ready to approve.

How a Slack mention becomes a market briefing email

2026

A Slack bot that turns a list of topics into a single coherent email briefing: it researches the news, correlates it with Futuur's prediction markets via LLM, then assembles a responsive HTML email and drafts it directly in SendGrid for review.

How a Slack mention becomes a market briefing email

Design, Dev & Product Loop

2026

Started as a Slack bot that triages bug reports into Trello, grew into a resilient E2E suite that diagnoses bugs against the live codebase, and now runs a design critique loop that evaluates its own UI proposals, remembers what it's learned, and opens pull requests for human review. Built solo end to end, except the PR-opening step, done alongside a developer.

NYC trip plannerPersonal2026

A PWA to plan a NYC trip: live map, curated events, and a drag and drop itinerary saved to Supabase. Conflict detection is deterministic — plain code correlating times, places, and distances — and the LLM is called only to explain what it found. Tokens go to the explanation, never to the detection. Built solo, full stack, deployed on AWS.

NYC trip planner

My Activity

@NariSake

369 contributions in 2026

Great products don't scale because of better interfaces. They scale because of better systems.

Wanna chat?

Even if this isn't about designing, I'm genuinely curious to hear your thoughts about AI, design and products in general.