Independent intelligence studio
A focused workspace for ambitious challenges, rigorous submissions, and work that earns its place.
Operations Console
Curated machine-learning quests across vision, language, audio, forecasting, and intelligent systems—each with a clear target and leaderboard.
Explore the questsEvery submission is compared with the solutions already submitted for its challenge.
| Challenge | Version | Real Solution ID | Submitted At | Status |
|---|---|---|---|---|
| No submissions logged yet. | ||||
Challenge rewards and weekly streak bonuses stack independently.
Place in the top three on 3 challenges for ₹5,000, or 5+ for ₹10,000.
| Source | Challenge | Rank | INR | Status |
|---|---|---|---|---|
| Loading earnings... | ||||
| Date | Challenge / Description | Type / Pool | Rank | Amount | Status |
|---|---|---|---|---|---|
| Loading payouts history... | |||||
Undo puts payouts back to their earlier status. Payouts changed since are left as they are.
| Intern / payment details | Challenge | Source | Rank | INR | Status | Action | |
|---|---|---|---|---|---|---|---|
| Loading payout ledger... | |||||||
Configure your settlement payout channels (UPI / Bank Account) and contact info.
| Username | Contact & Payout | Status | Credits (Used/Limit) | Base + Bonus | Actions |
|---|---|---|---|---|---|
| Loading intern accounts... | |||||
| Intern | Challenge | Routed Account | Bound Date |
|---|---|---|---|
| Loading challenge bindings... | |||
Credit a score the intern earned through their routed account, link that account's leaderboard entry to them, and optionally set their payout. Every change is immutable and attributed to the operator.
Review submissions whose remote outcome could not be confirmed. Verify the linked version before resolving an item.
epistra-account-link-helper.zip.chrome://extensions in Chrome.epistra-account-link-helper folder.manifest.json.
| Username | Role | Status | Action |
|---|---|---|---|
| Loading admin accounts... | |||
Submissions and audit findings whose solution.py matches earlier stored work. Open a case to compare both files before recording a decision.
| Case | Opened | Challenge | Intern | Matched | Kind | Method | Similarity | Status | Source |
|---|---|---|---|---|---|---|---|---|---|
Loading integrity cases... | |||||||||
Every solution.py checked at submission or read back from the evaluation accounts. Sort one challenge by score to read the top-ranked code.
Pick a challenge to sort by public or private score. Scores only compare within one challenge.
| Challenge | Intern | Account | Version | Result | Public score | Private score | Final rank | Uploaded | Open |
|---|---|---|---|---|---|---|---|---|---|
Loading stored solutions... | |||||||||
Stores every solution uploaded under the evaluation accounts in the chosen window and compares each with earlier uploads for the same challenge. Certain copies open audit cases for review; nothing is decided automatically.
| Version | Status | Score |
|---|---|---|
| No submissions recorded yet. | ||
| Rank | Solver | Score |
|---|---|---|
| No submissions recorded yet. | ||
import sys
from pathlib import Path
import pandas as pd
public_dir = Path(sys.argv[1])
submission_out = Path(sys.argv[2])
train = pd.read_csv(public_dir / "train.csv")
test = pd.read_csv(public_dir / "test.csv")
# Images, if present:
# image_path = public_dir / "train" / f"{image_id}.png"
# Train and predict...
submission_out.parent.mkdir(parents=True, exist_ok=True)
submission.to_csv(submission_out, index=False)