Daniel Ramos

Remoti · Santiago, Chile · Remote

Architected an enterprise RAG platform and a fleet of autonomous agents, then made both cheap enough and fast enough to run in production.

• Built end-to-end retrieval pipelines over messy enterprise data — PDFs, internal docs, logs — holding query responses under three seconds.
• Raised answer accuracy with dynamic chunking and hybrid search instead of naive top-k vector lookup.
• Designed stateful agents in LangGraph that use function calling with Pydantic schemas to query SQL databases and third-party REST APIs directly.
• Shipped async FastAPI microservices with SSE and WebSocket streaming, which cut time-to-first-token to something users read as instant.
• Ran LLMOps — token budgets, latency tracking and per-request cost — alongside prompt-security and data-privacy rules.

• Python
• FastAPI
• LangGraph
• LangChain
• OpenAI
• Gemini
• Qdrant
• pgvector
• Pydantic
• PostgreSQL
• Docker
• TypeScript
• React

Space-Eyes · Miami, Florida · Remote

Monitoring, security and defense systems — from sensor data processing up to the operator interface, across web, mobile and desktop.

• Led full-stack development on Morpheus, a counter-drone detection and response system, from the backend services through to the operator interface.
• Built SeaWatch, a maritime monitoring platform, and led its containerised deployment on AWS.
• Delivered FireWatch, a wildfire early-warning platform, including the React Native app that put alerts in the hands of field crews.
• Contributed to Mems, a companion detection product in the same family.
• Owned the backend spine across the product line — Python and FastAPI microservices, Docker containerisation, and deployment to both local and AWS environments.

• Python
• FastAPI
• React
• React Native
• TypeScript
• Docker
• AWS
• PostgreSQL
• Jira

Conalca SAS · Colombia · Remote

AI-assisted automation on Django, plus the cloud plumbing to keep it running without anyone babysitting it.

• Built automated, AI-driven internal tools in Python and Django to replace manual data work.
• Wrote scrapers and scheduled automations that turned scattered external sources into clean, queryable data.
• Administered the AWS footprint — EC2 instances, elastic IPs and domains — for the team's production deployments.
• Used AWS Glue to consolidate and move data between systems that had never spoken to each other.
• Maintained and improved the React and Tailwind CSS frontends, and coordinated a cross-functional team to keep releases on schedule.

• Python
• Django
• React
• Tailwind CSS
• AWS
• AWS Glue
• GCP
• Firebase
• SQL
• Web scraping

Independent · Remote clients · Remote

Custom Flutter apps for international clients, delivered solo — design handoff, backend, store build and all the glue in between.

• Shipped cross-platform apps in Flutter and Dart for clients across several countries.
• Ran the whole backend on Firebase — auth, Firestore, Cloud Functions and storage — with no separate server to maintain.
• Built RESTful APIs for integration with clients' existing internal platforms, verified with Postman.
• Added Firebase Cloud Messaging for real-time notifications, and Cloud Storage for uploads and downloads inside the apps.
• Worked from Figma designs and handled Android Studio setup, signing and release builds.

• Flutter
• Dart
• Firebase
• Cloud Functions
• GCP
• Figma
• Postman
• REST

Canal Capital · Bogotá, Colombia · Hybrid

Media content management for a public broadcaster, and the ERP work that taught me how much of engineering is other people's legacy code.

• Built web applications for managing and distributing multimedia content with Node.js and Express.
• Developed the frontend in React and TypeScript, factoring shared behaviour into custom hooks so features stopped being copy-pasted.
• Maintained and extended ERP systems written in PHP (MVC) on MySQL.
• Added Cloud Functions on Firebase and GCP for event-driven server-side work.

• Node.js
• Express
• React
• TypeScript
• PHP
• MySQL
• Firebase
• GCP