SOFTWARE ENGINEER · DEVOPS · BUENOS AIRES

PEDRO
ORTIZ DE GUINEA

Infrastructure & Trading Systems — front end to FIX session

I build software for algorithmic trading: the interface traders keep open all day, and the automation running behind it that they never have to look at. I joined Hybridge as part of the founding team and helped build the platform from inception alongside the rest of the team — FIX connectivity, back end, strategy logic, execution, dashboards, DevOps and automation. Today I am Product Owner of the front end and responsible for DevOps, infrastructure and version management.

If you can think it, I can program it.
Buenos Aires, Argentina — open to interesting conversations
Pedro Ortiz de Guinea
2+
years at Hybridge, from the founding team onward
20+
production environments running the platform
0→1
CI/CD designed and built from scratch
2
editions of the NOR derivatives program — I traded futures and options myself
// what i do

FRONT END & PRODUCT

I design and build trading interfaces, and I work as Product Owner for them: gathering what traders actually need, turning it into a roadmap, and shipping it. I rebuilt a functional but dated tool into a product people are comfortable sitting in front of for eight hours, following current front-end standards and UX practice borrowed from outside this industry.

Vue.js · JavaScript · WebSockets · real-time UI · UX · Product ownership

DEVOPS & INFRASTRUCTURE

I take care of everything between a commit and a running system: containerizing applications, building CI/CD pipelines, managing releases and versions across repositories, provisioning cloud infrastructure, and setting up the networking and access controls that regulated environments require.

Docker · GitHub Actions · Bitbucket Pipelines · AWS · Cloudflare Zero Trust · Nginx · Bash

LOW LEVEL & MARKETS

I am comfortable working close to the metal: FIX connectivity, market data and order routing, strategy logic and low-level components, C/C++ libraries, and Linux tuned for high-frequency and latency-sensitive workloads. I have also traded futures and options myself, which helps when deciding what a system should actually do.

FIX / QuickFIX · Python · C/C++ · ZMQ · Linux & kernel tuning · market microstructure
// the front end interactive — click around

Examples of the kind of interfaces I design and build, recreated here so you can actually use them: a dashboard that watches a whole desk, and the detail views for the two families of strategy — one that works through a quantity, and one that quotes both sides of a market.

monitor — live strategies CONNECTION
BYMA · logged in ROFEX · logged in MAE · reconnecting MD FEED · 12ms BACKUP · offline

PROGRESS STRATEGY · GGAL

RUNNING
BID4 810.00
OFFER4 815.00
LAST4 812.50
SPREAD5.00
FILLED
14 200
REMAINING
10 800
AVG PRICE
4 806.11
SLIPPAGE
-0.13%
PROGRESS57%
EDITABLE PARAMETERS — PUSHED TO THE RUNNING STRATEGY
STRATEGY FEED

        

MARKET MAKING STRATEGY · DLR/DIC25

RUNNING
BID1 441.00
OFFER1 444.00
LAST1 442.50
SPREAD3.00
POSITION
+320
P&L
+48 210
MY BID
1 441.50
MY OFFER
1 443.50
TRADES
184
EDITABLE PARAMETERS — PUSHED TO THE RUNNING STRATEGY
STRATEGY FEED

        
A note on these views: these are not screenshots — they are examples of how this kind of information gets presented. The real platform is deployed inside client networks, and it is not a finished, fixed product either: it is a development that adapts to specific uses and cases and changes with them. Everything above is a faithful recreation I hand-coded for this page, with invented data.

REAL-TIME WITHOUT THE JANK

These screens carry live market data and strategy state for a whole desk at once. Keeping that smooth is a front-end engineering problem in itself: batching updates, rendering only what is on screen, and making sure a busy session never costs the user a dropped frame or a frozen tab.

Vue.js · WebSockets · streaming state · virtual scrolling · performance profiling

BUILT WITH THE PEOPLE ON THE DESK

Everything here came out of working directly with traders: their feedback, their real use cases, and watching how they actually operate. On top of that I applied current front-end standards and technology. The result is meant to be adjusted — to a different desk, a different workflow, a different set of requirements — and as much of it automated as the user wants.

user research · iterative delivery · configurable views · role-based access · automation
// under the hood

The interface is the visible part. Most of my work is the engineering underneath it — and these are the areas I have worked in and can work in again.

DEVOPS & AUTOMATION

I design and run CI/CD from scratch: pipelines on GitHub Actions and Bitbucket Pipelines, self-hosted runners, container images built and published automatically, versioning and releases kept traceable across several repositories. I also automate what comes after the build — deployments, restarts, health checks, environment provisioning — because anything I have to do twice by hand is something I would rather write once.

Docker · GitHub Actions · Bitbucket Pipelines · self-hosted runners · AWS · ECR · Bash

SECURITY BACKGROUND

I started my career as a cybersecurity consultant at KPMG, working with organisations where the cost of getting it wrong is high: maturity assessments, cyber risk management, identifying vulnerabilities and building remediation plans with the client. I bring that lens to the systems I build now — thinking about attack surface, access control and secrets management while the thing is being designed rather than after.

cyber risk management · maturity assessments · hardening · zero-trust access · secrets management

NETWORKING

Getting an application reachable, encrypted and trusted in someone else's network is rarely the same problem twice. I have worked with reverse proxies, tunnels, certificate issuance and renewal, VPNs and DNS, and I am used to figuring out what will actually work given the constraints a particular client's environment imposes.

Nginx · Cloudflare Tunnels & Zero Trust · TLS / ACME · VPNs · DNS · firewalls

BACK END & LOW LEVEL

I work on the layer underneath the interface: FIX connectivity, market data and order routing, strategy logic and low-level components, messaging between processes, and the C/C++ libraries involved. I am comfortable reading and modifying code at that level, and comfortable in Linux when the answer turns out to be a kernel or system setting rather than the application.

Python · C/C++ · FIX / QuickFIX · ZMQ · asyncio · Linux internals

EFFICIENCY & SCALE

I spend real time on making software cheaper and steadier to run: profiling CPU and memory use, finding the hot path, cutting what a process wastes, and shrinking build artefacts and the transfer they cost on every release. It is unglamorous work with a direct payoff — the same hardware carries more, the system behaves predictably under load, and the bill goes down.

profiling · CPU & memory optimisation · concurrency · image size reduction · cost control

AGENTS & LLMs

I work closely with language models and build agents that do real jobs. I developed an automated code reviewer that reads every push, plus a set of agents and skills for internal automation and for documentation — turning technical material into something the business side can actually read, which shortens the distance between engineers and everyone else. I design the prompts and tooling, wire them into existing systems, and tune them against real usage.

LLM agents · automated code review · tool/skill design · prompt engineering · CI integration
// how i work

I studied computer engineering in Buenos Aires and joined Hybridge, in algorithmic trading technology, as part of the founding team before I had finished. A few things I have decided I believe:

$
I would rather delete a ticket than close it If something lands in my inbox twice, the third time should be a script. Many of the things I have built started as a chore somebody was doing by hand.
$
I build from feedback, not assumptions I work from what users tell me, from their real use cases, and from the standards of how this is done well elsewhere. I have watched traders work and I have traded myself, which makes those conversations shorter.
$
I work across the whole stack CSS one morning, a FIX session or a kernel setting that afternoon. I would rather understand a system end to end than be handed one layer of it, and most interesting problems sit on a boundary anyway.
$
Boring infrastructure is the goal Nobody should be able to tell that a release happened. If the deploy was exciting, I did it wrong.
// contact

THE ONE-PAGE VERSION

Same story, compressed onto a single sheet. Both versions are current.

CV — ENGLISH (PDF) CV — ESPAÑOL (PDF)