I connect commercial strategy, trusted analytics, forecasting and responsible AI to improve the way teams plan, review and act.
Career outcomes supplied in my professional brief, from separate initiatives. The public demos below contain educational and synthetic data; they do not independently verify these outcomes.
For commercial leaders, GTM Operations and Finance. Reconcile plan, actuals and forecast; review exceptions; approve a version and track the next action.
V83.2 — over 80 weekly production cycles. Includes AI-assisted regional narrative generation, called-vs-actual forecast accuracy tracking, and a structured approval workflow for global consolidation.
Try: compare a snapshot, save a call and export a WBR.
For RevOps, Sales, Marketing and GTM Strategy. Separate timing changes from demand, inspect conversion cohorts and test growth and resource alternatives.
Try: trace a push-out, compare acquisition economics and model capacity.
For analysts and leaders preparing commercial decisions. Retrieve shared metrics, inspect sources, test scope boundaries and retain human review.
Try: ask for coverage, change the region and inspect the returned evidence.
Production-grade queries used to build the commercial intelligence in this portfolio. Each answers a specific business question.
SELECT
f.fiscal_quarter,
f.week_of_quarter,
f.region,
f.forecast_value_usd AS called,
a.actual_bookings AS actual,
ROUND(
ABS(f.forecast_value_usd - a.actual_bookings)
/ NULLIF(a.actual_bookings, 0), 3
) AS error_pct,
CASE
WHEN f.forecast_value_usd > a.actual_bookings
THEN 'High'
WHEN f.forecast_value_usd < a.actual_bookings
THEN 'Low'
ELSE 'Accurate'
END AS bias
FROM weekly_forecast_log f
JOIN quarter_actuals a
USING (fiscal_quarter, region)
WHERE f.forecast_type = 'Commit'
ORDER BY 1 DESC, 2, 3SELECT
segment,
fiscal_quarter,
COUNT(DISTINCT opportunity_id) AS num_opps,
ROUND(AVG(arr_usd), 0) AS avg_acv,
ROUND(
SUM(CASE WHEN outcome='Won'
THEN 1.0 ELSE 0 END)
/ NULLIF(COUNT(*), 0), 3
) AS win_rate,
ROUND(AVG(sales_cycle_days), 0) AS avg_days,
ROUND(
COUNT(DISTINCT opportunity_id)
* AVG(arr_usd)
* (SUM(CASE WHEN outcome='Won'
THEN 1.0 ELSE 0 END)
/ NULLIF(COUNT(*), 0))
/ NULLIF(AVG(sales_cycle_days), 0), 0
) AS velocity_per_day
FROM closed_opportunities
WHERE close_date >= CURRENT_DATE - INTERVAL '365 days'
GROUP BY 1, 2
ORDER BY 2 DESC, velocity_per_day DESCStart from the decision, align definitions, test competing explanations and assign the follow-up. SQL and automation make that mechanism reproducible.
The Kaggle CRM source keeps its original 2016–2017 dates. Planning, subscriptions and payments use independent deterministic scenarios. Missing facts remain visible.
python scripts/reproduce.py python -m http.server 8000
My work spans sales performance, incentive governance, forecasting, analytics, data quality and global enablement. I build operating mechanisms across teams and work hands-on with SQL, Power BI, Python, Salesforce and enterprise data platforms.
This independent implementation demonstrates analytical and operating-model patterns. It contains no employer code or records. The AI integration follows the production pattern I have used with large language models for automated narrative generation, anomaly detection and commercial summaries — applied here to Kaggle CRM and independently generated planning data.