All projects

Education

Spar app icon

Spar

Interview Sparring

Spar is a local desktop gym for data structures and algorithms. Open a LeetCode-style workspace, write Python or JavaScript, and run tests on your machine. A 12-week path, streaks, and XP keep the daily loop going — with no hearts, so you can fail Submit as often as you need. After a problem is Accepted, an interview-style AI review can come from OpenAI or a local Ollama model. Progress stays in SQLite on disk, not in someone else’s cloud.

Download Spar
Back to portfolio
Spar desktop app — home and interview workspace on a dark gold stage

01

Daily gym

Home opens on the next problem, not a dashboard of chores. Streak, XP today, and accepted count sit above today’s prompt — Continue drops you straight into the workspace. Daily goals track accepted solutions. There are no hearts: you can run and submit until the tests go green.

Spar home — streak, XP, accepted count, and today’s problem with Continue

02

A 12-week path

Twelve units in the order they build on each other — arrays, trees, graphs, DP. Finish a unit’s core problems to unlock the next. Stretch problems never block you. The path is medium-first interview work, not a random problem feed.

Spar path — twelve units with core and stretch interview problems

03

Interview workspace

Full problem statement on the left, Monaco editor on the right. Switch Python or JavaScript, run visible tests, submit hidden ones. Drafts autosave per language. Notes, editorials after Accepted, a timer, and a submissions history sit in the same window so the loop stays on one screen.

Spar workspace — Contains Duplicate with editor, tests, Run, and Submit

04

Local judge

Run and Submit execute with the Python and Node runtimes on your machine. Visible cases teach; hidden cases decide Accepted. If python3 or node is missing, setup asks for an absolute path instead of failing silently.

05

AI interview review

Once a solution is Accepted, review it as an interview coach would — complexity, edge cases to mention, a cleaner approach — without a full rewrite. Use an OpenAI key or a local Ollama model from Settings. Reviews persist in SQLite keyed by problem and language, so coming back still shows the last one.

Spar AI review — complexity, edge cases, and a cleaner approach after Accepted

06

Progress that stays on device

Drafts, notes, streaks, XP, submissions, and reviews live in a local SQLite database. Optional AI credentials never leave the machine except to the provider you chose. Installers ship as a signed macOS Apple Silicon .dmg and Windows .exe / .msi.

Spar progress — six-week activity and accepted counts by pattern

Get Spar

Built and shipped by Diwakar IT Services as a real product — not a mock showcase.

  • LeetCode-style workspace in Python and JavaScript
  • 12-week path with streaks, XP, and unlimited submits
  • Local judge with visible and hidden tests
  • AI review via OpenAI or a local Ollama model
Download Spar
Ask us about building something like this

More products

Healthcare

Niti

A secure, compliance-minded clinic operations product for front desk and care teams — patients, appointments, admissions, and prescriptions with auditability built in.

Fintech / Productivity

Timmy

A group expense app for shared spending — splitting costs, tracking balances, and keeping money conversations simple.

Security

Opty

An offline two-factor authenticator for moments when connectivity is unreliable — security that still works without the network.

Product engineering

moklay

A managed URL shortener for teams and developers — organize branded short links in-app, then connect systems with a public Integration API and scoped keys.