CV

  1. Software Engineer

    Sep 2024 – Present
    Snowflake

    Working on Snowpipe Streaming, a low-latency streaming ingestion and transformation service.

    • Core contributor to the data and control planes, from code complete to GA.
    • Owned the Kafka Connector, schema evolution, Iceberg integration.
    • Mentored an intern on the implementation of distributed tracing and fine-grained metrics.
    • Improving the team's engineering culture: introduced stacked PRs, Java records and sealed interfaces, pytest fixtures, Mermaid diagrams.
    • Solved infrastructural issues around networking and containers.
    • Spearheaded the use of AI for investigations.
    full-timeDBMSC/C++PythonJavaAI

  2. Software Engineer & Tech Lead

    Sep 2021 – Jul 2024
    Firebolt

    Helped build Firebolt's nervous system—a distributed, transactional metadata layer.

    • Led RBAC development for Firebolt 2.0 (5 people, 3 months).
    • Led the engine-database separation project (2-5 people, 6 months).
    • Owned the System Engine, a centrepiece microservice for Firebolt's vision of a SQL-only API.
    • In 2022, initiated the design and implementation of a new Logical Query Plan. It's more capable and orders of magnitude faster than its predecessor.
    • Implemented benchmarking, algorithmic optimization, and observability of performance.
    • Contributed major improvements to architecture, code quality, and developer productivity.
    • Stack: C++20, Bash, Python, Go, SQL, GitHub Actions, Docker.
    full-timeDBMSC/C++Python

  3. Internship & Master's Thesis

    Oct 2020 – Jul 2021
    SAP

    SAP HANA group, Walldorf. Implemented a partitioned hash join operator on an FPGA using Intel OneAPI.

    DBMSC/C++low level

  4. Internship & Semester Project

    Feb 2020 – Aug 2020
    IBM Research Zurich

    Mathematical optimization, CPU & GPU implementations, and benchmarking of ridge regression.

    • Came up with a novel, faster formula for solving ridge regression with intercept in closed form. Wrote up a 4-page convex optimization convergence proof and implemented it in C++ using eigen, OpenMP and CUDA.
    • It outperformed scikit-learn and NVIDIA RAPIDS by orders of magnitude. It was quickly deployed to an IBM product used by financial institutions and contributed to a NeurIPS 2020 publication: https://arxiv.org/abs/2006.09745.
    • Also implemented a GBM tree-building algorithm on a GPU with state-of-the-art performance.
    • Stack: C++, eigen, OpenMP, CUDA, LaTeX.
    C/C++AIlow level

  5. MSc Information Technology & Electrical Engineering

    Sep 2018 – Jul 2021
    ETH Zurich

    MSc Information Technology + Electrical Engineering, GPA 5.51.

    • Advanced computer architecture, GPUs, FPGAs, data management systems, big data, HPC.
    • Machine learning, probability, signal processing, information theory.
    • Thesis: "Acceleration of DBMS query execution on FPGAs", advised by prof. Gustavo Alonso.
    educationDBMSAI

  6. Co-founder

    Jan 2018 – Jan 2024
    Sol

    Co-founded Sol to help preserve the identity of rare musical instruments. We supplied secure NFC tags with an accompanying digital service to serve as an instrument's digital passport.

    • Management: marketing, sales, accounting, legal.
    • Front-end: web and mobile apps, landing page.
    • Cryptography: tools for provisioning and simulating NFC tags using proprietary protocols from NXP.
    • Stack: Python, TypeScript, React, HTML + CSS, Docker.
    Pythonfull stack

  7. Software Engineer

    Jun 2016 – Jul 2017
    Cambridge Consultants

    Worked on multiple embedded projects ranging from satellite and aviation transceivers to machine vision.

    • Very low-level software: using L1 cache as SRAM, custom bootloaders, handling PCIe interrupts, multi-core profiling and synchronization, software cache coherence, atomic DMA queues.
    • Python tooling: parsing ELF files, running CI on embedded hardware, automating lab experiments.
    • Played a vital role in a major project for our highest-revenue customer.
    • Fixed complex critical production bugs affecting multi-million-dollar telecoms product lines.
    • Had a customer request that I personally work on their new project.
    • Found undocumented silicon bugs in SoCs.
    • Technologies: C, Python, MATLAB, assembly, TCP/IP, OpenCV, Bluetooth, VHDL, Cadence Allegro.
    full-timeC/C++Pythonlow level

  8. BEng Electronic Engineering

    Sep 2014 – Jun 2018
    University of Manchester

    1st class. Scholarship from the UK Electronics Skills Foundation.

    education

  9. Embedded Systems Engineer

    Jun 2014 – Aug 2014
    Internet InterActive Promotion

    Reverse-engineered a schematic diagram and various communication protocols, and later developed new firmware for an existing non-documented ARM-based cycle rental station. Served 20,000 users.

    • Technologies: C++, I2C, SPI, RS485, Modbus.
    C/C++low level