Learn to Achieve — Empowering the learners public systems leave behind

The through-line

We've been doing the same job since 2017. The tools just caught up. L2A operating thesis

01.

Information should be free.

Public data, plainly built. If a tool needs a login, a paywall, or a six-figure contract to access, we built the wrong thing. Everything ships open-source on GitHub, MIT-licensed by default.

02.

Tools, not gatekeepers.

A student choosing a college, an overworked agency analyst, a patient picking a hospital — each hits a decision point where the data they need is locked behind expertise they don't have. We unlock it.

03.

Continuity, not pivots.

We've built tools for under-resourced education and public institutions since 2017. The mediums change — school-data tooling then, civic data now. The mission doesn't.

04.

With public partners, not against them.

State agencies, school districts, and county health offices aren't our adversaries. They're underfunded, understaffed, and over-deadlined. We build tools that make their work easier — not reports that make them look bad.

Why now

The ladder broke. The data's locked. We work the gap.

The entry-level white-collar ladder has collapsed — automation absorbing first-rung tasks, work moving offshore, a frozen hiring market. At the same moment, the public data that could actually help sits scattered across agencies, unusable. Learn to Achieve works the intersection: build the public-interest tools that need building, with the people who need the work.

5.7%

recent-grad unemployment — now above the national rate

41.5%

of recent grads underemployed — working jobs that don't need their degree

2017

building tools for under-resourced institutions since

100%

of what we build ships open-source and free to use

Labor figures: Federal Reserve Bank of New York, Q1 2026.

What we build

We don't just talk about shipping. We ship.

Everything here is open-source and built in the open — the same real, verifiable work the platform exists to create.

98 federal datasets

17.7M+ rows

27 agencies

MIT licensed

Open-source platform

Lighthouse Open Health

An open-source platform that makes the full corpus of federally-disclosed U.S. healthcare data queryable in one place — a living systematic review with an AI analyst sitting on top of the raw data, not pre-computed summaries. Three lenses on one dataset:

● Live

01 · Workforce

CA Workforce Atlas

Maps where California's clinician supply falls short of patient need — joining state physician supply, preventable-hospitalization rates, social vulnerability, and federal shortage areas on county FIPS. Anchor: Glenn County, 45 physicians per 100K vs. 334 statewide.

02 · Accountability

Provider record

One place to see a facility's full federal accountability record — CMS deficiencies, malpractice and adverse actions, exclusions, and fraud recoveries — keyed on a single facility ID. Collapses a four-agency search into one query.

03 · Outbreak

Outbreak Watch

A rolling situation report tying the CDC's current health advisories to the live national surveillance feeds behind them — respiratory, vector-borne, and foodborne. The featured advisory rotates as new alerts publish.

Python · Streamlit · DuckDB · Cloudflare R2 (Parquet lakehouse) · Neon Postgres · Plotly · multi-provider LLM RAG (~$0.003/query)

View on GitHub →

grindIn

◐ In development

L2A's first product — a job-discovery and application engine. A daily pipeline pulls 10,000+ postings from 8 sources, integrates with major applicant-tracking systems, and scores and ranks every role. 437+ tests, full CI/CD.

View on GitHub →

Triagent

◐ In development

An agentic triage trainer built on live GitHub issues. An LLM agent pulls real open issues, scores them for solvability and difficulty, and ranks them into a work-ready queue — so early-career engineers practice on real bugs, not toy problems.

View on GitHub →

Workforce GIS

○ In scoping

A public dashboard mapping healthcare-worker shortages county by county across California — replacing slow manual GIS work with an open, validated pipeline. In scoping with state workforce and health-data partners.

Partner brief — in progress

How it works

Real work, made to count.

The model we're building: connect people to real open-source and public-interest work, then turn what they do into something an employer recognizes.

Pick

Choose a real project — an open-source tool or an operational problem at a public partner.

Build

Do the actual work, in the open, with a verifiable trail on GitHub.

Verify

The partner confirms the impact. A public, hosted case study, not a self-report.

Translate

Turn the work into the economic-value language hiring managers actually recognize.

Bank

Stack each case into a portable portfolio that travels with you.

This is the system we're building — the tools above are the proof the model works. We're onboarding our first cohort and partners now.

Why it's different

Real work, open access, an economic-value translation, and a public portfolio — nobody runs all four together.

Bootcamps

Projects are simulated, with no third-party verification.

Volunteer boards

Matching only — no credential layer, no employer bridge.

Pro-bono networks

Serve nonprofits with people who are already credentialed; closed to learners.

University capstones

Sometimes real, but gated to the enrolled, and the portfolio doesn't travel.

Our story

How we got here.

A straight line from a small education nonprofit to the civic-tech work we do now — same mission, sharper tools.

2017

Founded in far-northern California

Learn to Achieve started as a public benefit nonprofit with one aim — help students reach education and careers beyond high school.

2018

The first data tool

We built a reporting system on a local school district's student-information data, turning messy records into something usable. The data-engineering work that defines L2A today started right here.

2019 – 2021

Education programs on the north coast

We ran education and college-prep programming across the region, working with the students and schools around us.

2022 – 2025

A quiet stretch

The organization paused while its people went off to build and learn elsewhere — at Google and across the wider tech and data world.

2026

Relaunched as civic-tech

We came back with a sharpened mission: free, open-source public-interest data tools, and a real way into the workforce for people the entry-level market has shut out.

The team

Small team. Big range.

A small group of builders who've shipped at Google and across federal health data — plus the people backing and steering where it all goes.

Sahil Amresh Ratnam

Founder · since 2017

Founded Learn to Achieve in far-northern California. A software engineer and data scientist by trade, actively involved in the relaunch.

LinkedIn ↗

Roberto Morales

Co-founder · since 2017

Built L2A's engineering from the ground up. A senior AI backend engineer who deployed Gemini across 10M+ devices at Google Nest; co-leads the org's direction.

LinkedIn ↗

Joseph Abad

Director of Workforce Programs · 2026

A data analyst at the California Department of Health Care Access & Information. Helps steer the organization's direction and its workforce-data partnerships.

LinkedIn ↗

Venura Wijenayake

Technical Director · since 2018

A CS-grad data engineer who runs the technical side. Built Lighthouse and grindIn — Python and SQL pipelines, REST APIs, and production RAG that real users depend on.

LinkedIn ↗

Get involved

Three ways in.

We're onboarding our first cohort and partners now — early is the best time to get in.

Build something real

Recent grad, switcher, or anyone shut out of the first job — pick a project, do the work, and walk away with a verified case study.

Start building →

Partner with us

A nonprofit, school district, or agency with a data or operational problem you can't afford to fix? Bring it and get skilled help.

Become a partner →

Support the work

We're a California 501(c)(3) nonprofit. Help us put more public-interest tools into the world and more people into careers.

Online giving — coming soon