Devansh Panwar

Create your portfolio with ProoV
ProoV
DE Verified work portfolio

ProoV Portfolio

Devansh Panwar

Artificial intelligence and machine learning · Graphic era deemed to be university

View the ProoV leaderboard

Projects

Certified

Self-directed project

Energy and Renewable: Machine Learning for the Power Grid

Junior Machine Learning Engineer · AI / ML · September 2026

84/ 100

A ProoV case study · educational project, not employment

Built a leakage-aware turbine yield predictor end to end: framed TEY as a regression target, excluded downstream CO/NOx emissions, trained both a LinearRegression baseline and a RandomForest model, and validated on a genuine time-based holdout. The final report includes held-out MAE/RMSE/R², a residual read, and a practical monitoring plan for drift and retraining.

Graded against

  • Build and validate the model correctly35%
  • Optimize and evaluate honestly25%
  • Write the Model Card (clarity + honesty)25%
  • Explain it to a non-expert15%

Passed · pass mark 60/100

What stood out6
  • Correctly excluded CO and NOx as downstream leakage features and documented why
  • Used an earlier-period train / later-period test split instead of a random shuffle
  • Compared a linear baseline against a tree ensemble and showed a real held-out MAE improvement
  • Read residual behavior and noted that metrics do not prove unbiased performance across all operating ranges
  • Leakage-Free Feature Selection
  • Time-Based Model Validation
The work I submitted20 tasks

12 tasks · 13 lines · Python, JavaScript · 8 written answers · 437 words

  • Problem Frame180 characters
  • Leakage Flags113 characters
  • Train Validate BaselinePython · 1.4k characters
  • Train Validate Forest942 characters
  • Train Validate124 characters
  • Tune Evaluate SetupPython · 1.5k characters
  • Tune Evaluate ResidualsPython · 2 lines
  • Tune Evaluate129 characters
  • Residual Read247 characters
  • Drift Monitor92 characters
  • Model Card994 characters
  • Teach BackJavaScript · 738 characters
  • Problem Frame15 words
  • Leakage Flags1 word
  • Train Validate Baseline141 words
  • Train Validate Forest76 words
  • Train Validate1 word
  • Tune Evaluate Setup110 words
  • Tune Evaluate Residuals92 words
  • Tune Evaluate1 word

Summarised on purpose — the submitted code and writing stay private so this page cannot be reused as an answer key. The full submission sits behind the verified certificate.

Verified certificateTamper-proof · issued by ProoV
Certified

Self-directed project

AI-Assisted Code Modernization with IBM

Junior Software Engineer · Software Engineering · August 2026

84/ 100

Based on a public IBM case study · not affiliated with IBM

Built and verified a full legacy-code modernization for KM-Waechter: corrected the service-wear calculation, made missing readings safe, fixed the nightly report’s average and crash path, and caught a hidden mileage-conversion bug in the helper layer. Also added the missing regression test, produced a data-driven breakdown-risk analysis from fleet_history.csv, and validated the result with an acceptance script before handing back the GitHub repo.

Graded against

  • Fix and extend the repo with the agent35%
  • Diagnose the bug20%
  • Direct and audit the AI agent25%
  • Explain it like an engineer20%

Passed · pass mark 60/100

What stood out6
  • Fixed the nearly-worn-car bug with true division so 14,900/15,000 now evaluates to about 99.3% instead of 0%
  • Handled missing last-service readings without crashing the report and without falsely flagging the car
  • Found and corrected the quiet fleet_utils.km_to_miles conversion bug that no test directly caught
  • Built a data-driven breakdown analysis that explicitly identifies km_since_service, avg_daily_km, and load_factor as the real signals
  • Leakage-Free Feature Engineering
  • Honest Model Evaluation
The work I submitted20 tasks

12 tasks · 1.6k characters · 8 written answers · 78 words

  • Division Predict72 characters
  • Diagnose Line130 characters
  • Diagnose Why95 characters
  • Scope Prompt66 characters
  • Audit Findings158 characters
  • Audit Why167 characters
  • Acceptance Checklist100 characters
  • Template Repo Click82 characters
  • Bob Trial Click324 characters
  • Exec Pitch209 characters
  • Bob Bonus174 characters
  • Verify Report71 characters
  • Division Predict1 word
  • Diagnose Line10 words
  • Diagnose Why17 words
  • Scope Prompt1 word
  • Audit Findings1 word
  • Audit Why29 words
  • Acceptance Checklist18 words
  • Template Repo Click1 word

Summarised on purpose — the submitted code and writing stay private so this page cannot be reused as an answer key. The full submission sits behind the verified certificate.

Verified certificateTamper-proof · issued by ProoV
2Projects completed
2Verified certificates
84Average score
Create your portfolio with ProoV