Sanya Ray — Portfolio

Projects

Certified

Self-directed project

AI-Assisted Code Modernization with IBM

Junior Software Engineer · Software Engineering · September 2026

96/ 100

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

Built a complete modernization and bug-fix pass for the KM-Waechter service: corrected the wear math, handled missing service readings safely, fixed the nightly report’s average and crash path, and repaired a hidden mileage-conversion bug in the helpers. Also added the missing regression test, produced a data-driven breakdown-risk analysis from fleet_history.csv, and verified the repo against all 11 acceptance checks before handing it in.

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

Caught and fixed the hidden fleet_utils.km_to_miles conversion bug that no test directly covered.
Preserved the required 15000 km interval and 80% threshold while fixing the wear calculation and missing-reading behavior.
Added the missing regression test for a car with no reading and verified the full checklist with verify.py.
Wrote an analysis that explicitly rejects the misleading odometer/age assumption and ranks risk using the real predictors.
Leakage-Free Bug Diagnosis
Directed AI Code Review

The deliverable

View the codeSanyaR2908 / proov-km-waechter-fix

The work I submitted20 tasks

12 tasks · 2.4k characters · 8 written answers · 97 words

Division Predict75 characters
Diagnose Line130 characters
Diagnose Why151 characters
Scope Prompt66 characters
Audit Findings158 characters
Audit Why168 characters
Acceptance Checklist214 characters
Template Repo Click82 characters
Bob Trial Click384 characters
Bob Ready60 characters
Verify Report70 characters
Exec Pitch794 characters
Division Predict1 word
Diagnose Line10 words
Diagnose Why25 words
Scope Prompt1 word
Audit Findings1 word
Audit Why24 words
Acceptance Checklist34 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

Certified

Self-directed project

Amazon Robotics - Black-Friday Fleet Coordination

Software Engineering · August 2026

96/ 100

A ProoV case study · educational project, not employment

Built a full Black-Friday fleet coordination solution end to end: modeled the warehouse as a typed grid, implemented optimal single-agent A* with Manhattan heuristic, extended it to cooperative multi-agent planning with a centralized reservation table that prevents both vertex and edge collisions, and selected realistic charging/idle policies to keep the fleet moving. In simulation, the candidate identified the throughput sweet spot, achieved zero collisions and zero deadlocks, and packaged the result into an executive-ready pilot recommendation.

Graded against

Problem framing & operational stakes20%
Algorithmic depth - A* and cooperative coordination30%
Coordination policy realism - deadlock, charging, idle25%
Throughput trade-off analysis15%
Executive pitch to leadership10%

Passed · pass mark 55/100

What stood out6

Implemented A* with an admissible Manhattan heuristic on a 4-connected grid and verified it beat Dijkstra in node expansions.
Built a reservation table keyed by cell and tick, and explicitly prevented head-on swaps with directed edge reservations.
Chose a realistic fleet policy: centralized coordination, priority/watchdog deadlock handling, staggered charging, and dispersed perimeter idle parking.
Used simulator outputs to lock onto the concave throughput optimum at 35 robots and justified the choice with wait-time and collision data.
Leakage-Free Cooperative Planning
Honest Simulator-Based Evaluation

The work I submitted20 tasks

12 tasks · 6.4k characters · 8 written answers · 248 words

Role And Mission115 characters
Amazon Robotics Context459 characters
Grid World Theory119 characters
State Representation Task341 characters
Phase1 Recap348 characters
Astar Theory124 characters
Astar Implementation Cell3.8k characters
Astar Solution Correct4 characters
Single Unit Challenge465 characters
Phase2 Recap374 characters
Multi Agent Problem111 characters
Reservation Table Theory180 characters
Role And Mission1 word
Amazon Robotics Context37 words
Grid World Theory1 word
State Representation Task1 word
Phase1 Recap23 words
Astar Theory1 word
Astar Implementation Cell171 words
Single Unit Challenge13 words

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

Machine Learning for Automotive

AI / ML · July 2026

78/ 100

A ProoV case study · educational project, not employment

Built an end-to-end used-car pricing analysis for VW and Audi listings: combined the datasets, explored price and mileage patterns, and compared Linear Regression with Random Forest using holdout evaluation. The work also translated model performance into a business-facing pricing exposure estimate and recommended a data-driven pricing workflow for the CPO team.

Graded against

Exploratory Data Analysis20%
Feature Engineering25%
Model Building & Evaluation35%
Business Communication20%

Passed · pass mark 60/100

What stood out6

Combined VW and Audi listings into a single analysis table with a brand indicator and used it consistently across EDA.
Used multiple EDA views: histograms, correlation heatmap, scatter plots, and a brand boxplot to compare market behavior.
Identified the strongest business drivers in the model output and translated MAE into a pricing exposure estimate for the CPO team.
Compared Linear Regression against Random Forest and reported both R²-style accuracy and MAE differences in the business summary.
Exploratory Data Analysis
Leakage-Free Feature Preparation

The work I submitted10 tasks

5 tasks · 6 lines · 5 written answers · 196 words

VW Sprint1 Complete56 characters
VW Sprint2 Complete61 characters
VW Sprint3 Complete59 characters
VW Business Insights1.4k characters
Jupyterlite Code2 lines
VW Sprint1 Complete1 word
VW Sprint2 Complete1 word
VW Sprint3 Complete1 word
VW Business Insights192 words
Jupyterlite Code1 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