Manuel Martín Morante · Engineer & researcher
I build software and explore the mathematics behind it, from database infrastructure and scientific computing to generative models for drug design.
ScrollThe work
Engineering, research, and entrepreneurship. Several roles overlapped.
Now
Founder & Lead Engineer · Philadelphia, PA · On-site
Building the next generation of peptide therapeutics alongside William Herbert, Prof. César de la Fuente and Prof. Marcelo Der Torossian Torres, as a spin-off of UPenn.
Since Oct 2026
to
Software Engineer
Developed ClickHouse and core database infrastructure, and worked on platform infrastructure at scale with Kubernetes and Docker. Rebuilt the billing system into a cloud-provider-grade model, working across teams on correctness and reliability.
Madrid, Spain · On-site
to
Emergent Venture Fellow · George Mason University
Pursued independent research in protein design, supported by two Emergent Ventures grants.
Two research grants of 4K each.
to
Undergraduate Researcher
Built an end-to-end generative pipeline for therapeutic peptides targeting GSK3β, a kinase involved in tau hyperphosphorylation in Alzheimer’s disease. Combined molecular generation, target-specific guidance, and a blood–brain barrier permeability classifier.
to
Co-founder & Advisor
Co-founded and advised a company building marketing mix models for large ecommerce businesses using causal machine learning.
The company was sold before the product launched.
to
Undergraduate Researcher · Valencian Research Institute for Artificial Intelligence
Worked on computational genomics and precision medicine. Co-authored “Exploring BioNER Frontiers: An In-Depth Evaluation” (CAiSE 2025), a systematic benchmark of biomedical named entity recognition models.
Valencia, Spain · On-site
to
Undergraduate Researcher
Conducted exploratory research at the intersection of AI, drug design, and biotechnology, including time in Rafael Gómez-Bombarelli’s lab. Worked with diffusion models and flow matching for molecular generation and guidance, alongside GPU-level implementation to run these methods efficiently at scale.
Later recognition: MIT Grand Hackathon winner, March 2026.
to
Co-founder · Head of Data & AI
Co-founded a B2B SaaS used by Meliá and PcComponentes, and led its Data & AI division. Built the full data and ML stack from scratch with ClickHouse, Python, Rust, and PyTorch, including scraping infrastructure handling millions of requests per day. Owned hiring and growth end-to-end.
Raised €500K and grew the company to $1M ARR.
Remote
to
SEO Consultant & Web Developer
Worked with companies including Mercado Libre and ABCDIN on data-driven SEO and product marketing. Built and deployed custom ecommerce applications and production APIs, connecting marketing analysis with hands-on software development.
Full-time
Foundations Data Science at UPV Valencia and Mathematics at VIU. Theory and practice, side by side.
The questions
Different ways of looking at the same world.
The language underneath it all. I’m interested in the ideas behind the algorithms, from linear algebra and probability to the geometry of learning.
A place to slow down and question my assumptions. I write about ambition, agency, and what it means to choose a direction in life.
Understanding things by building them. Database engines, GPU kernels, molecular simulations: I like following an idea all the way down to the implementation.
Another way of understanding the world. I’m interested in how art expresses ideas, challenges familiar perspectives, and gives us something new to think about.
Surfing and snowboarding, alongside Muay Thai, BJJ, and powerlifting. A way to get out of my head, stay present, and test my limits.
Experimenting away from the keyboard. I enjoy the process of turning simple ingredients into something worth sharing.
Field notes
Technical deep dives and more personal lines of thought.
A practical deep dive into Rotary Positional Embeddings (RoPE) in CUDA: mathematical derivation, tensor layouts, coalesced and vectorized memory access, and effective bandwidth optimization on modern GPUs.
A practical deep dive into physical merge algorithms in MergeTree: why ClickHouse uses horizontal and vertical merges, what trade-offs they optimize, and what we learn from building a hybrid adaptive chunked merge in C++.
Hace una semana firmé mi sueño de mudarme a USA a montar una empresa. Reflexiones sobre el sacrificio, el coste personal y la realidad tras alcanzar una meta.
Agency, ownership y todo eso: en el idioma de mi abuela era echarle pelotas. Nadie te regala nada; hay que tocar todas las puertas.