Yarco Hayduk
Early-Stage Web3 Venture • Distributed Systems • Quantitative Research

Yarco Hayduk, PhD

General Partner at Pragma Ventures, a research-oriented Web3 fund. PhD in distributed systems. Since 2021 leading investments end to end, from sourcing and technical diligence through final decisions, across blockchain infrastructure, stablecoins, tokenization, and AI-native systems, co-investing alongside a16z, Coinbase Ventures, Archetype, and Variant. In crypto since 2017.

Based between Lisbon and Seoul. Citizenship: Ukraine & Canada. US TN-eligible.
Expertise
Early-Stage Venture Technical Due Diligence Digital Assets Stablecoins, RWA & Tokenization DeFi Web3 x AI Blockchain Infrastructure Distributed Systems Algorithmic Trading
Publications

Research across distributed systems, concurrent computing, blockchain systems, and lock-free order book designs.

DAIS · OPODIS · ManLang · NETYS · DATE · DASFAA · DSD
Experience
06/2021 – Present
General Partner & Technical Lead
Pragma Ventures pragma.ooo Lisbon / Seoul

A private, research-oriented Web3 fund focused on blockchain infrastructure, tokenization, stablecoins, and AI-native systems.

  • Ran the investment process end to end, from founder interviews and deep reviews of whitepapers, protocol design, and systems architecture through internal investment memos and final decisions.
  • Deployed checks of $100k–$750k across 20+ portfolio companies to date.
  • Sourced and closed recent deals including OpenGradient, a verifiable AI inference network (round led by a16z, alongside Coinbase Ventures), TACEO, a private execution layer for finance (seed led by Archetype, with a16z CSX participating), and Exo, distributed AI inference on consumer hardware, backed ahead of its open-source release (44k+ GitHub stars today).
  • Built a market intelligence system tracking public activity from 7,000+ crypto investors and builders on X, catching investable themes early and reading where early-stage consensus is heading, alongside AI-assisted diligence workflows.
  • Supported portfolio companies on systems challenges spanning indexing, parallel execution, infrastructure reliability, and on-chain data.
  • Founding member of an invite-only Web3 investor community in Lisbon and co-founder of a local agentic engineering community. Judge at Web3 hackathons and conferences.
11/2020 – 02/2021
Senior Blockchain Engineer
CovalentVancouver, Canada

Blockchain indexing infrastructure spanning multiple EVM networks with differing reorg and finality behavior.

  • Oversaw the design of a decentralized indexing pipeline, retrofitting a local multi-chain EVM indexing system backed by relational databases for deployment on the Moonbeam Network.
  • Refined multi-chain ingestion, consistency, and query paths to preserve the reliability and performance of the centralized system across supported EVM networks.
  • Contributed to strategies for handling reorgs, finality, and chain-specific indexing behavior across supported EVM networks.
Stack:
ElixirErlangPythonJavaSQLPostgreSQL
06/2019 – 09/2020
Quantitative Engineer
Hex CapitalVancouver, Canada

A proprietary trading firm building quantitative strategies and high-frequency trading infrastructure across crypto markets, with live multi-venue execution spanning both DeFi and CeFi.

  • Drove the design and implementation of proprietary market-neutral strategies and high-frequency trading systems across crypto venues, with a strong focus on execution quality, arbitrage capture, and execution risk management.
  • Built resilient trading infrastructure spanning exchange connectivity, order routing, monitoring, and failure handling for live multi-venue deployment.
  • Developed the observability stack for real-time latency profiling, execution validation, and strategy performance analysis.
Stack:
GoPythonPrometheusPostgreSQL
03/2019 – 06/2019
Polyglot Developer
Bluzelle NetworksVancouver, Canada

Decentralized censorship-resistant storage network with multi-language client libraries.

  • Built and maintained the Go, C#, Ruby, and C++ client libraries, preserving strict protocol consistency across runtimes.
Stack:
GoC#RubyC++
11/2017 – 03/2019
Principal Developer
HQS ConsultingWinnipeg, Canada

Enterprise software and infrastructure consulting across JVM systems, cloud platforms, and production web applications.

  • Designed and built enterprise applications in Scala and Java.
  • Built backend services and web platforms using Spring-based stacks, relational databases, and containerized deployment workflows for client-facing products.
  • Owned deployment and operational environments on GCP, covering infrastructure administration, monitoring, release workflows, and ongoing production maintenance.
Stack:
ScalaJavaSpringDockerGCP
10/2012 – 04/2017
Lead Research Engineer
ParaDIME Project, EU FP7 Switzerland / Belgium / Germany

EU-funded consortium (Neuchâtel, TU Dresden, BSC Barcelona, IMEC Belgium) targeting energy optimization across the full data center stack, with production integration work for IMEC Belgium and Cloud & Heat Technologies in Germany. This work was conducted alongside the PhD and produced most of the resulting publications.

  • Designed a distributed message-passing framework on top of Akka (Actor Model) to accelerate execution across available CPU, GPU, and distributed resources.
  • Developed a compile-time code rewriting framework using Scala Macros for transparent energy-aware optimizations, requiring no application-level changes.
  • Integrated agent-based coordination into existing systems at IMEC Belgium and Cloud & Heat Technologies to support seamless production operations.
  • Received a Best Paper Award at DAIS 2016 for work on energy-efficient actor execution on heterogeneous architectures.
Stack:
JavaScalaAkkaCLLVMCUDASparkASM
01/2009 – 12/2009
Java Developer
Zilliant via SoftServeRemote

B2B pricing optimization and revenue management software for large product catalogs, pricing workflows, and data-driven commercial decision-making.

  • Contributed to Java-based pricing optimization software used to support large-scale catalog pricing, price guidance, and margin-focused decision workflows.
  • Worked on data-intensive application components involving pricing logic, statistical analysis, and high-volume product and customer datasets.
  • Overhauled ORM query batching and related data-access paths to improve data-fetching throughput by 40%.
Stack:
JavaSpringHibernateREST
Education
11/2017 – 07/2018
Postdoctoral Researcher
University of British ColumbiaVancouver, Canada
Research on frequent pattern mining for Big Data streams using modern stream processing frameworks and heterogeneous CPU/GPU execution.
Faculty: Faculty of Engineering
Group: Computer and Software Systems
Advisor: Prof. Alexandra (Sasha) Fedorova
Focus: High-throughput stream analytics, latency and throughput benchmarking, and concurrent CPU/GPU execution.
  • Proposed and benchmarked streaming data-mining workflows in Spark, Flink, and Storm, with a focus on latency and throughput in real-world settings.
  • Explored concurrent versions of FP-streaming algorithms designed to exploit multi-core CPUs and GPU resources for high-throughput stream analytics.
  • Applied the work to social data mining and large-scale event streams, including evaluation of production-style workloads on cloud infrastructure.
Stack:
SparkFlinkStormScalaCUDA
10/2012 – 07/2017
PhD, Computer Science
Université de NeuchâtelNeuchâtel, Switzerland
Research spanning concurrency, energy-efficient computing, Big Data systems, GPU computing, and systems-level performance engineering.
Group: Complex Systems and Big Data Lab
Advisor: Prof. Pascal Felber
Thesis: Exploiting Concurrency and Heterogeneity for Energy-efficient Computing: An Actor-based Approach.
  • Built a heterogeneous task scheduler routing work between GPU and CPU resources based on real-time energy cost, reducing consumption by 40–80%.
  • Designed a lock-free concurrent order book for HFT workloads, compatible with LMAX Disruptor, with 15x speedup on a 16-core machine.
  • Modified OpenJDK HotSpot C++ to enable pauseless garbage collection in ParallelNew and CMS using Hardware Transactional Memory.
  • Consulted on concurrency workflow design for robotic arm motion control at Innovation Mining & Incorporated Projects Inc.
TA: Concurrency, Security, Cloud Computing (2012–2016)
01/2010 – 05/2012
MSc, Computer Science
University of ManitobaWinnipeg, Canada
Research focused on distributed data mining and uncertain-data pattern mining using MapReduce and Hadoop.
Group: Database and Data Mining Lab
Advisor: Prof. Carson Leung
Thesis: Distributed frequent pattern mining algorithms for uncertain data with MapReduce.
GPA: 4.38/4.5
  • Implemented frequent pattern mining algorithms for regular and uncertain data on Apache Mahout and Hadoop.
  • Used MapReduce and ForkJoin on Amazon EC2 to achieve 9.5x speedup on 11 nodes.
TA: Analysis of Algorithms, Automata Theory, Database Concepts
09/2009 – 06/2012
MSc, Computer Science
Lviv Polytechnic National UniversityLviv, Ukraine
Research focused on parallel frequent pattern mining for Big Data.
Topic: Parallel frequent pattern mining for Big Data.
09/2005 – 06/2009
BSc, Computer Science
Lviv Polytechnic National UniversityLviv, Ukraine
Studies in computer science with an early focus on concurrency middleware and systems programming.
Topic: Comparison of concurrency middleware in C# and Java.
Result: GPA 4.91/5, 2nd out of 95.
Selected Publications
DAIS
Enhanced Energy Efficiency with the Actor Model on Heterogeneous Architectures
Y. Hayduk, A. Sobe, P. Felber
Best Paper Award
NETYS
Exploiting Concurrency in Domain-Specific Data Structures: A Concurrent Order Book and Workload Generator for Online Trading
R. P. Barazzutti, Y. Hayduk, P. Felber, E. Rivière
ManLang
Towards an Efficient Pauseless Java GC with Selective HTM-Based Access Barriers
M. Carpen-Amarie, Y. Hayduk, P. Felber, C. Fetzer, G. Thomas, D. Dice

Full list at dblp.org