# Pharedata

> Pharedata designs, builds and deploys AI agents, machine learning systems and data engineering for companies that compete on their data. Based in Santiago, Chile; works in Spanish and English. Systems go to production with observability, controlled costs and outcomes tied to the business — not endless pilots.

- Website: https://pharedata.com/ (Spanish by default, English available through the language selector)
- Contact: contacto@pharedata.com
- LinkedIn: https://www.linkedin.com/company/pharedata/
- Location: Santiago, Chile (founded 2017)
- Sectors: retail, e-commerce, finance, industry, SaaS
- Clients shown on the site: Copec, Latam, Paris, Easy, Esval, Sotraser

## Capabilities

- [Autonomous AI agents](https://pharedata.com/#capacidades): agents that operate on a company's systems — research, decide and execute back-office tasks with controlled tooling and guardrails. Enterprise RAG, multi-agent systems, evaluation.
- [Applied machine learning](https://pharedata.com/#capacidades): predictive and recommendation models trained on the client's data — forecasting, churn, pricing, scoring and anomaly detection.
- [Data engineering](https://pharedata.com/#capacidades): pipelines, lakehouses and data models built to feed AI, with quality, governance, observability and cost control from day one.
- [Cloud automation](https://pharedata.com/#capacidades): critical processes automated on AWS, GCP or Azure, including integrations, to remove repetitive manual work and speed up decisions.

## How we work

- [Process](https://pharedata.com/#proceso): short sprints, each ending in something demonstrable.
  1. Discovery (1–2 weeks): understand the business, map the available data, prioritize use cases by impact and feasibility.
  2. Design (2–4 weeks): technical architecture, cost estimates, success metrics and deployment plan.
  3. Implementation (4–10 weeks): pipelines, models and agents, delivered as a production version, not a prototype.
  4. Iteration (ongoing): monitoring, continuous improvement, model evaluation and new capabilities.

## Use cases

- [Retail and e-commerce](https://pharedata.com/#casos): an operations agent connected to catalog, inventory and ad spend that detects sales drops, identifies causes and proposes actions to the commercial team every morning.
- [Finance](https://pharedata.com/#casos): near real-time pipelines feeding customer scoring models and atypical-transaction detection, integrated with the existing financial core.
- [Industry and B2B](https://pharedata.com/#casos): RAG assistants over technical documentation, manuals and historical reports that answer with citations for operations, support and sales teams.
- [SaaS and product](https://pharedata.com/#casos): forecasting and recommendation models delivered as APIs, with drift monitoring and automated retraining.

## Technology

- [Stack](https://pharedata.com/#stack): models and agents — OpenAI, Anthropic, LangChain, LangGraph, LlamaIndex, PyTorch; data and platform — AWS, GCP, Azure, Databricks, Snowflake, BigQuery; pipelines and apps — Python, dbt, Airflow, FastAPI, Next.js, Postgres.

## Team

- [Team](https://pharedata.com/#equipo): more than 20 years of experience in strategy consulting, engineering and data science.
  - Felipe — Analytics · ML. MBA, University of Edinburgh; Master in Information & Data Science, UC Berkeley; AWS Certified Solutions Architect; 15+ years in advanced analytics.
  - Guido — Technology · Data. Computer Engineer, PUC Chile; in technology since 2003, leading engineering teams; focused on turning data and AI into business value.
  - Ana María — Engineering · Product. Computer Science Engineer, PUC Chile; AWS Certified Solutions Architect; 15+ years in backend, mobile and AWS architecture; builds production AI agents and RAG systems.

## Contact

- [Book a call](https://pharedata.com/#contacto): the first conversation is free and without commitment — contacto@pharedata.com