Photo of Adrian Castro

Adrian Castro

Backend Tech Lead at Specter · Distributed systems, data platforms, APIs

Based in München, Germany | Italian (C2), Spanish (C2), English (C1), German (B1), French (A2), Portuguese (A2)

Adrian Castro

Backend Tech Lead at Specter · Distributed systems, data platforms, APIs

Based in München, Germany | Italian (C2), Spanish (C2), English (C1), German (B1), French (A2), Portuguese (A2)

M.Sc. Data Engineering @ TUM, 2019 – 2022

B.Sc. Computer Science @ UniMiB, 2016 – 2019

Experience

Specter

Feb 2026 – Present
  1. Backend Tech Lead Sep 2026 – Present
    • Built the company's MCP server from scratch and led the design and rollout of most new public APIs.
    • Design and build distributed systems that process large volumes of data reliably and securely.
    • Lead architectural decisions across the platform, focusing on scalability and performance under high load.
    • Drive cross-functional initiatives with Product, Design and Engineering, and help shape the technical roadmap with leadership.
    • Mentor engineers through design reviews, pair programming and knowledge sharing.
    • Bringing good vibes and good food
  2. Senior Software Engineer Feb 2026 – Aug 2026

Senior Software Engineer @ Tacto Technology GmbH

Sep 2023 – Dec 2025
  • Mostly focused on the Analytics product, which is responsible for driving 60% of Tacto's ARR, per time of writing.
  • Designed and shipped end-to-end features in a multi-tenant SaaS platform, from data pipelines and APIs to customer-facing dashboards.
  • Improved performance, reliability and observability of core services, and took part in code reviews, on-call rotations and technical planning.

Data Engineer @ IONITY GmbH

Sep 2022 – Aug 2023
  • Led the modernization of the entire data platform from the ground up on AWS, reducing costs by 50% with a focus on high availability, scalability and security.
  • Built Airflow, Glue and Kinesis Firehose pipelines handling hundreds of thousands of records per day from tens of sources, speeding up existing pipelines by 50% to 250%.
  • Introduced software-engineering practices, GitLab CI/CD for Terraform and company-wide data governance standards, drastically reducing technical debt.

Technical Co-Founder @ Joleat

2019 – 2022
  • Pitched our company to hundreds of people
  • Grew the company from 0 to 400 sales in just 6 months
  • Onboarded 60+ business partners to sell on our platform
  • Everything with no capital invested in marketing
Software Engineer @ D4L data4life gGmbH Sep 2021 – Aug 2022
Software Engineer @ Heveloon Ltd Oct 2020 – Sep 2022
Staff Board Member @ UnixMiB 2017 – Present
Software Engineer @ BioDataAnalysis GmbH 2020 – Aug 2021
Full-Stack Developer @ Net2Web Srl 2017 – 2021
Internship @ Università degli Studi di Milano-Bicocca Mar 2019 – Jun 2019
Junior Partner @ Bicocca Security Laboratory 2017 – 2019

Technologies and Knowledge

Languages: Python, JavaScript / Node.js, C++, SQL

Cloud & infrastructure: AWS, Azure, Terraform, Kubernetes, CI/CD, Linux / Unix

Data: Data engineering, Airflow, batch and event processing, databases

Architecture: Distributed systems, microservices, REST APIs, MCP, high performance computing

Other: Deep learning, DevOps, agile development

Events

Linux Day Milano 2019 - Event dedicated to Open Source, GNU/Linux, free software, sharing and open culture
Akademy 2019 - Organized the yearly KDE Akademy with 200+ people with UnixMiB
MLConf San Francisco 2019 - Helped organize the yearly MLConf in SF
FOSDEM - Participant since 2019

Papers and Essays

(TUM) Is This Real Life? - An essay about the possible future implications of AI onto society.
(TUM) Exploring and Analyzing Compression Algorithms in Preprocessing Pipelines - A report about my work extending and testing TensorFlow with Zstandard on Machine Learning preprocessing pipelines.
(TUM) Model Parallelism: Challenges, State-of-the-art, and Future Directions - A survey of current state-of-the-art in model parallelism.

Tools

(UniMiB) PSO for Cancer Evolution - The aim for the script is to find the optimal cancer evolution tree given an input matrix representing cancer cells
(TUM) WP Collect Implicit Feedback - A WordPress plugin that tracks user clicks and path across the website, with the goal of feeding the data to interactive recommender systems.
CDash CLI Client - CDash API wrapper in Python.

Hobbies

I love to cook every kind of food and dish. I am a big fan of South American cuisine, especially Ecuadorian and Peruvian. I also enjoy kickboxing and piano, although the latter is more like a dead hobby.
My weakness is pausing movies or TV series to explain something happening there.

References

Available on request