About

I’m looking for a data scientist who can build an automated valuation model similar to Zillow/Redfin/Realtor estimates. You’ll be working with large, messy property datasets and turning them into a clean, research-driven valuation system. This isn’t a comps job. I need someone who approaches valuation analytically, tests assumptions, and can explain their reasoning clearly. What You’ll Do -Work with ~1.2M property records from multiple real-estate sources -Merge, clean, and normalize inconsistent datasets - Identify which property attributes and neighborhood factors drive value - Run statistical analysis to detect patterns, biases, and anomalies - Build fair-value estimates, success-probability predictions, and expected-reduction models - Document your methodology so the engineering team can implement it easily - Extract simple, actionable insights from the data What You Should Be Good At - Strong statistical foundation (regression, hypothesis testing, distribution analysis) - Python or R (pandas, NumPy, SciPy, statsmodels, scikit-learn, XGBoost/LightGBM, etc) - Working with large and messy datasets - Independent research and analytical thinking - Experience with real-estate, AVMs, or tax-assessment data is a bonus Deliverables - Clean consolidated dataset - Three working models (value estimate, success probability, expected reduction) - Clear documentation of preprocessing and inference steps - Summary of insights found in the data