CTO and AI technical leader · Mathematical Engineer (EAFIT) · Medellín, Colombia
I lead teams that build AI products and software, with a background in applied mathematics. Over a decade of experience: from founding a software company (Arista Dev, since 2013) to working as an external CTO for portfolio companies.
📫 LinkedIn · ✉️ estebanhiguita@gmail.com · 📍 Medellín, Colombia (UTC-5)
- AI for finance and product. I lead the development and rollout of AI solutions that improve processes and decision-making.
- Technical leadership. I have led multidisciplinary teams, including a core team of 8 developers at a legal-AI startup, and defined engineering practices as a Lead Architect.
- Applied mathematics. Actuarial and statistical modeling; research with CEPAR, UNSW Business School (2016–2018) and EAFIT.
- Immersive tech. Augmented reality and 3D: Unity, Vuforia, WebAR, Three.js.
- CTO, Winter Kpital (2024–present): external CTO for portfolio companies, technical review of investments, AI for finance.
- Founder and CTO, Arista Dev (2013–present).
- Co-founder and Technical Advisor, After Co (2022–present).
- Before: Technical Lead and AI Specialist at Legal AI Latam (2023–2024), Lead Architect at Workep (2020–2021).
| Repo | What it is |
|---|---|
| MortalityProjectionInterface | Shiny app for mortality projection with the Lee-Carter, CBD, APC, M7 and RH models (R, StMoMo). |
| gpt4-pdf-chatbot-langchain | My fork of mayooear/ai-pdf-chatbot-langchain: a GPT-4 and LangChain chatbot for large PDFs, which I used to study the pattern. |
| GestionSystem | Dart web app for administrable content (2015). |
Most of my recent work is private client and portfolio code, so it is not listed here. Happy to talk architecture and AI — reach out on LinkedIn.
AI / ML: Python · TensorFlow / Keras · PyTorch · scikit-learn Software: C# · JavaScript · Python · Java · PHP Data and math: R · Shiny · statistical and actuarial modeling Immersive: Unity 3D · Vuforia · WebAR · Three.js · Oculus
- CEPAR, UNSW Business School (Sydney), 2016–2018: a model that lets users without programming experience build data-analysis models for actuarial companies.
- Logic and Computation Research Group, EAFIT (2015–2017): a paper on the Black-Scholes equation applied to drug absorption, presented at a conference in Baltimore, USA.
English / Spanish (native).
