Ashley Karani

About Me

I'm Ashley. I don't just build tools — I build decision systems.

I operate at the intersection of data, strategy, and business performance, partnering with teams to transform evolving data into a strategic asset that drives revenue, reduces risk, and enables faster, higher-quality decision-making.

I focus on decision intelligence. I design and scale data ecosystems that connect engineering, analytics, and governance into a unified layer users can trust — moving organizations from reactive analysis to proactive, insight-led execution.

I measure success in outcomes:

Revenue growth through better allocation decisions
Cost efficiency through optimized data architecture
Risk reduction through strong governance and visibility

My role is to ensure that every critical decision is backed by clarity, not assumption — and that data becomes a competitive advantage embedded in how the business operates.

My Work

01

AI Engineering

Claude-Backed Internal Agent

Internal chat and agent surface that answers data questions over warehouse context. Semantic layer as retrieval surface, Claude tool use to hit live dbt models, prompt caching to keep per-request cost flat.

Tools Anthropic SDK, Tool Use, Prompt Caching, MCP, Semantic Layer

Production tool, used daily by analytics + ops

02

Data Engineering & Analytics

Customer Data Platform (CDP) & Unified Customer View

Customer data was fragmented across CRM, digital, and transactional systems, limiting personalization and analytics. Built a CDP integrating multi-source data into a single Golden Record customer profile with identity resolution and behavioral enrichment.

Tools Snowflake, dbt, SQL, Identity Resolution, CRM Integrations

Segmentation, churn analysis & personalized engagement at scale

03

Data Engineering

Near-Real-Time Operations

Rebuilt 4-hour batch pipelines into dynamic tables with ~1-minute latency. Replaced fragile materialized views with incremental dbt models, SLA-backed lineage, and observability on every mart.

Tools Snowflake Dynamic Tables, dbt, Event Hub, Airflow

4-hour batches → 60-second freshness

04

Data Engineering · Full-Stack

Ops Alert Platform

End-to-end data product pushing condition-triggered alerts to operational teams. React SPA + Azure Functions API + scheduled Snowflake queries + Graph API email delivery, all orchestrated through Logic Apps.

Tools Azure Functions, Static Web Apps, Snowflake, Graph API

Manual watchlist → zero-touch alerting

05

Analytics Engineering

Inventory Aging Decision Layer

Unified item-vendor model across 10+ ERPs with a decision layer exposing dead-inventory risk scores. Partnered with sales to prioritize older stock before it aged out.

Tools Snowflake, dbt, Data Contracts, Power BI

~60% reduction in dead inventory

06

Analytics Engineering

Canonical Yield Models

Unified yield definitions across multiple plants and ERPs. Standardized transformation logic, automated historical snapshots, and exposed a semantic layer so analysts stopped re-deriving the same numbers five different ways.

Tools dbt, Snowflake, Medallion, Semantic Layer

~20% improvement in yield accuracy

My career &
experience

Delivered weekly, decision-ready dashboards and reports for fitness, media, and retail clients — giving multiple departments a shared view of performance they could act on.
Built and maintained ETL pipelines across outsourced applications, CRMs, and multi-customer source systems — improving data reliability and cutting manual reconciliation.
Designed a RAG pipeline POC for a retail client using open-source LLMs and vector search over product catalogs — targeting a 15% lift in customer engagement through intelligent product discovery.

Turned complex datasets into executive-ready reports and dashboards — increasing stakeholder clarity and adoption of key performance insights.
Led Mastercard Foundation data controls and reporting — establishing governance and visibility over critical program metrics.
Automated manual operational workflows, driving a 23% gain in efficiency and accuracy across business processes.
Managed end-to-end data storage and ETL (Hevo, Stitch) — keeping data flows organized and accelerating ad-hoc analysis turnaround.

Owned a global data pipeline (Airbyte, Stitch) integrating Snowflake, CRM, and third-party sources — unifying fragmented data for cross-region reporting.
Engineered in-database cleaning and transformation workflows — raising data quality and reducing time-to-analysis for downstream teams.
Built Metabase, Power BI, and Google Data Studio views that translated complex datasets into actionable business narratives.
Supported product and marketing analytics via Mixpanel and Google Analytics — including funnel analysis that informed product evaluation and growth decisions.

Conducted primary and secondary research on growth rates, consumer trends, and market potential — grounding product and expansion strategy in evidence.
Stood up the Market Intelligence data function from scratch — delivering dynamic reporting infrastructure that supported executive strategic planning.
Presented findings to leadership and cross-functional teams — aligning the organization on market conditions and high-opportunity growth areas.

40%

Faster Reporting Cycles

3×

Pipeline Throughput

50+

Dashboards Delivered

$2M+

Cost Savings Identified

End-to-End

The Data Lifecycle

Click a layer to explore.

Collection & Data Ingestion

Technologies

Depth over logos

What I run in practice.

📥

Ingest & Store

SnowflakeAirbyte HevoStitch AWS S3Redshift Glue

⚡

Model & Transform

dbtSQL PythonSpark Airflow

📊

Visualize & Serve

Power BITableau MetabaseLooker Google Data StudioMixpanel Google Analytics

🤖

AI & Analytics

RAGLLMs Vector DBscikit-learn PandasNumPy

Principles

My Data Philosophy

Outcomes Over Outputs

I measure success by business impact, not lines of code. Every pipeline, model, and dashboard should answer a real question.

Data as Product

Treat data assets like products: documented, tested, versioned, and user-focused. Quality and reliability are foundational.

Simplicity Through Layers

Complex problems decompose into simple layers. Staging to intermediate to marts — each with a clear purpose and contract.

Contact

Let's work together

Whether you have an opportunity, a project idea, or just want to connect — my inbox is always open.

Email

ashleykarani6@gmail.com

LinkedIn

Location

Nairobi, Kenya

Say Hello