Sales Operations & Compensation — Align Technology
Automation · Analytics · Governed AI
I build practical automation, analytics and governed AI solutions that reduce manual work, improve compensation data quality, and give business teams faster access to reliable operational insights.
I'm a Sales Operations & Compensation Analyst (Level 5) at Align Technology, with a background that spans Biomedical Engineering, Business Administration and a Global MBA completed in 2025.
Over the past few years I've built end-to-end solutions for compensation operations: Python scripts that generate compensation documents at scale, governed AI agents that handle intake and routing inside Teams and Salesforce, Power BI embedded tools for live writeback, and SQL validation jobs that act as data quality gates before payouts run.
Before Align I worked as an Azure Cloud Support Engineer at Tek Experts (Microsoft) and as a Sales & Marketing Specialist at Medsys, where I designed CRM operating models and built the company's first virtual lead-capture chatbot.
Governed AI agent for the North America compensation team. Handles intake classification, internal research, case note generation and routing — grounded in approved data sources with audit-ready guardrails.
Business Problem
Analysts spent significant time triaging compensation requests — reading emails, cross-referencing payout data across multiple systems and drafting internal notes before beginning to resolve disputes. No systematic classification or pre-analysis existed.
Solution Built
Designed and architected SONA: a governed AI agent that classifies incoming requests by type (dispute, quota, crediting, policy), selects the correct approved data source, retrieves relevant context and generates structured internal case notes — reducing analyst prep work per case.
Technical Architecture
Copilot Studio (orchestration) → Azure OpenAI (NLU + generation) → SharePoint KB (approved policy docs) → Starburst SQL (compensation data) → Salesforce (case creation) → surfaced in Microsoft Teams.
Governance Design
Every response grounded in approved sources only. No invented payout amounts, policy rules or case owners. Missing identifiers trigger clarification. No supported source routes to human review. Source attribution in every note.
Stack
Key Features
Request classification by category · Governed source routing · Structured case note generation · AI guardrails (no hallucinated data) · Salesforce case creation · Teams-native interface · Escalation to human review
Business Impact: Reduced analyst time on pre-analysis and case note drafting per incoming request. Improved consistency of internal documentation. Established a governed AI pattern reusable for future compensation workflows with full audit trail.
Interactive tool embedded inside Power BI for submitting quota adjustments from a live dashboard. SharePoint writeback + Starburst DirectQuery drives real-time DAX recalculation on every submission.
Business Problem
Quota refinement inputs were collected through email threads and disconnected spreadsheets. No audit trail. Analysts couldn't see the real-time impact of adjustments on quota or attainment figures without manual recalculation.
Solution Built
A Power Apps form embedded directly inside a Power BI report. Users submit adjustments from the same interface they use to analyze data. Inputs persist to a governed SharePoint list, feeding back into the DirectQuery model for live DAX recalculation.
DAX Logic
Updated Quota = [Original Quota] + COALESCE(SUM(Refinement[InputValue]), 0)
QoQ change uses safe division for zero-base periods. Attainment % recalculates against updated quota in real time across all visuals.
Power Fx Upsert Pattern
Checks for existing record on (TerritoryID + QuotaType). If exists → updates. If not → creates new. User email and timestamp stamped automatically. Reset pattern for correction workflows.
Stack
Key Features
Upsert logic (create or update) · User-stamped submissions · Reset/delete for corrections · Live DAX recalculation · Python validation sync job · Reusable writeback pattern
Business Impact: Replaced disconnected email-based input collection with a governed, traceable system inside the analytics tool teams already use. Adjustments are auditable with full user and timestamp trail. Impact visible in real time without switching context.
Python pipeline that generates Sales Incentive Plan documents at scale. Reads role, country, OTI and payout curve data from a master workbook, assembles modular Word templates, inserts dynamic payout tables and outputs print-ready .docx files per role.
Business Problem
Producing SIP documents for a multi-role, multi-country plan required assembling content from multiple sources, inserting payout tables manually and applying role-specific formatting — time-consuming, inconsistent and error-prone.
Solution Built
Python pipeline reads a master Excel workbook (Summary + Payout Curves), iterates every role-country combination, assembles cover + custom section + T&C + component Word templates, replaces placeholders at run level, inserts dynamic payout tables and saves a final .docx per role.
Key Technical Logic
Curve key: role_bucket + " - " + component_name · Column normalization: strip/lower/replace on ingest · Run-level placeholder replacement (preserves Word formatting) · President's Club: Σ(attainment × weight) · Quarter-changes variant via config flag
Stack
Key Features
Batch generation for all roles and countries · Modular document assembly · Formatting-safe placeholder replacement · Dynamic payout table insertion · President's Club weighted attainment · Quarter-changes variant via config flag
Impact
Reduced document generation from a multi-hour manual process to a single script run. Eliminated manual copy-paste errors in payout tables and OTI figures. Allows regeneration with updated data at any point with zero additional effort.
Business Impact: Reduced a multi-hour manual document production process to a single script run. Eliminated formatting inconsistencies and copy-paste errors. Team can regenerate all documents at any point in the cycle with updated data and zero additional effort.
Automated validation comparing manual compensation adjustments against downstream actuals — flagging records that are missing, incorrectly credited or mismatched before they cause payout errors.
Business Problem
Manual adjustments entered in compensation systems don't always propagate correctly to downstream attainment actuals. Without a systematic check, discrepancies could go undetected until the payout cycle was complete — requiring costly corrections after the fact.
Solution Built
SQL mismatch detection query using CTEs to compare every manual adjustment against reflected actuals, classifying each as Matched, Missing In Actuals or Credit Mismatch. Python layer runs Pandas merge, applies vectorized status logic and exports a timestamped exception report.
SQL Pattern
Left join: manual adjustments → actuals on (order_id, territory_id, quota_type). CASE logic: NULL = Missing · delta ≠ 0 = Credit Mismatch · else Matched. CTE structure separates source prep from comparison for full auditability.
Stack
Key Features
Three-way classification · Multi-key join · CTE-based SQL · Python exception report · Designed as pre-payout data quality gate · Reproducible per cycle
Business Impact: Provides a repeatable pre-payout data quality check that catches discrepancies before they reach the payout stage. Reduces risk of incorrect compensation payments and manual effort required to identify and correct post-payout issues.
Power BI operational dashboard for tracking compensation disputes and tickets — open cases, resolution time, aging and case mix in real time. Replaced manual status consolidation with a governed, always-current source of truth.
Business Problem
Compensation ticket status was tracked through emails and spreadsheets. Leadership lacked a consistent view of open cases, case aging or team throughput. Status updates required manual consolidation before every team meeting.
Solution Built
Power BI dashboard connected to compensation case data, featuring open/closed KPIs, aging metrics, case type breakdowns and analyst workload views. All aggregation in DAX measures to keep the model clean and fully sliceable.
DAX Measures
Open Tickets = CALCULATE([Ticket Count], Status <> "Closed")Ticket Age = DATEDIFF(MIN([Created Date]), TODAY(), DAY)
Closed rate, avg resolution time, case mix — all filter-context aware.
Stack
Key Features
Open vs closed KPIs · Aging analysis by status · Case type and priority breakdown · Analyst workload view · Validated slicer interactions · Relationship-aware filter context
Business Impact: Gave the team a live operational view replacing manual status consolidation. Leadership reviews ticket health before every meeting without analyst prep time. Case aging visibility enables earlier intervention on stalled disputes.
Governed AI agent for intake classification, internal research and case note generation in Microsoft Teams.
Teams assistant letting analysts run approved compensation query presets using natural language prompts.
Integration design connecting ADA to SONA for downstream case note generation and analyst handoff via Salesforce.
Python mailer detecting compensation triggers in email, querying the approved knowledge base and generating governed responses.
Batch Python pipeline generating SIP documents for all roles and countries from a master Excel workbook using modular Word templates.
Quarter-change variant using a config flag to switch workbook sources without modifying core generation logic.
Power BI-embedded Power Apps form for quota adjustment submissions with live DAX recalculation via SharePoint writeback.
Reusable architecture for embedding business input forms in Power BI with governed SharePoint persistence and downstream sync.
Power BI dashboard for tracking compensation disputes — open cases, aging, case mix and workload in real time.
Python engine validating prorated payouts using attainment, payout curve lookup, OTI and active-day proration.
SQL + Python validation flagging manual adjustments not correctly reflected in downstream actuals before payout runs.
Multi-sheet Excel automation reading adjustment trackers with varying column structures, normalizing and comparing against actuals.
Python + MSAL pipeline authenticating, locating and downloading Excel files from SharePoint without OneDrive sync dependency.
Python + MSAL script triggering scheduled dataset refreshes via Power BI REST API and service principal — no manual clicks.
Standardized compensation data model: role buckets, component names, curve keys, OTI structures and payout table format.
Executive documentation of SONA operating modes with governance design, data sources, security and business flow.
Process design routing email compensation requests into Salesforce cases with record type, queue assignment and SONA trigger.
DirectQuery architecture connecting Power BI to Starburst for live compensation reporting with DAX quota, attainment and adjustment deltas.
Open to conversations about automation, analytics, AI systems or roles that push this work further. Reach out directly — I respond within 24 hours.