MSc in Artificial Intelligence
Universidad de los Andes
Team thesis project at the FLAG Lab: natural language processing of Spanish-language scientific documents, covering rhetorical segmentation and contribution extraction.
CamiloDurango
I connect business, data, and applied AI, and turn technical capability into decisions.
I do business and customer intelligence for Twilio’s customer-experience organization, from Medellín, Colombia.
I came up through finance, moved into data, and now build LLM-based systems that run in production.
Finance, then data, then applied AI
I started out in finance. I took a business undergraduate degree, then a master of international business at Hult International Business School in San Francisco, and worked as a financial analyst. Those years left me with the idea that still organizes my work: a number is only worth something when somebody can make a decision with it.
From there I moved into data, and then into applied AI. Today I work on business and customer intelligence for Twilio’s customer-experience organization, where I design and run LLM-based classification pipelines that turn unstructured customer interactions into structured data teams can report on. In 2026 I finished a master of science in artificial intelligence at Universidad de los Andes.
I spent several years in the San Francisco Bay Area and then chose to come back to Colombia. What I do best is explain technical ideas to people who are not technical, and connect what a system can genuinely do to a business problem worth solving.
Where I work
Business Intelligence Analyst
TwilioCustomer Experience Operations
Business and customer intelligence for the customer-experience organization: analytics and measurement of the customer experience, and systems that turn text into data.
- Analytics and measurement of the customer experience for the teams that run it.
- Design and operation of LLM-based classification pipelines in production, turning unstructured customer interactions into structured data for reporting and decisions.
- Monitoring and evaluation of those pipelines: distribution-drift checks on the predicted classes, and LLM-as-a-judge review.
- Recurring briefs that summarize escalations for the people who act on them.
- I have taught prompting strategies to mostly non-technical colleagues, and presented internally on using AI to automate text-heavy workflows.
Education and credentials
Applications of AI for Anomaly Detection
Master of International Business
Business undergraduate degree
What I work on
Business intelligence and analytics
Level of experience: Professional experience
Measurement, reporting, and data models teams use to decide something, not to file away.
Applied AI and LLM-based systems
Level of experience: Professional experience
Pipelines in production that pull structure out of text, with monitoring and evaluation around them.
Machine learning and data science
Level of experience: Graduate training and applied work
A master of science in artificial intelligence, and natural language processing work in Spanish.
AI-enabled workflows and automation
Level of experience: Practice at work, and an area I am exploring
Automating text-heavy work, and working out how it applies to small and mid-sized companies.
What I am thinking about
Making AI genuinely useful inside a business
Fewer demos, more processes that still run after the meeting ends.
Helping non-technical people understand AI
What it can do today, what it cannot, and how to judge an output without being an engineer.
Automation for small and mid-sized businesses in Colombia
Much of the repetitive work in these companies is text, which is exactly where there is room to help.
Building systems by hand
Writing the pieces myself, without a framework hiding them, so I understand how they work.
Let us talk
Write to me if you want to talk about business intelligence, applied AI, or automating text-heavy work. I read everything that comes in.