Building high-performance trading infrastructure and intelligent AI systems at the intersection of financial markets, low-latency computing, and large language models. CTO at two companies. Open-source author with 70k+ downloads. $56M+ in trading volume.
About
I'm Bertin Balouki SIMYELI, a quantitative developer, systems engineer, and AI architect with roots in West Africa. I build institutional-grade trading infrastructure and intelligent AI systems at the intersection of performance and correctness.
Since 2020, I've operated as a Business Partner and Introducing Broker with JustMarkets, driving over $56 million in trading volume across 316+ standard lots. This real-world market exposure directly informs the systems I engineer, from HFT parsers to sentiment-driven signal generation.
My technical practice spans Python & C++20 for high-performance trading systems, and LlamaIndex, RAG pipelines, and multi-agent AI for intelligent enterprise automation. I've led engineering teams as CTO at two companies and built open-source tools trusted by traders worldwide.
I run TheFinTech Insights Digest, a newsletter exploring distributed systems, fintech infrastructure, and the evolving landscape of algorithmic markets. I also offer professional consulting services for custom trading systems, risk models, and AI data pipelines.
Machine Intelligence
Architecting intelligent systems, from RAG pipelines and LLM integrations to reinforcement learning and multi-agent orchestration, applied to real-world financial and enterprise use cases.
AI is not just a skill set, it's the lens through which I approach modern systems engineering. From architecting RAG pipelines that automate business diagnostics to deploying multi-modal LLM agents for enterprise SaaS platforms, I bridge the gap between cutting-edge research and production-grade deployment.
My AI work spans financial markets (sentiment-driven trading signals, reinforcement learning strategies), edtech (Agent-Based Cognition systems), and enterprise automation (document AI, intelligent booking, growth analytics).
Forward Outlook Ltd: Architected a specialized Retrieval-Augmented Generation system using LlamaIndex to automate "Growth Gap" diagnostics, synthesizing complex multi-source business data into structured, actionable intelligence for the BGA networking community.
Movable Games: Leading development of a neuroscience-backed educational platform using Agent-Based Cognition, mathematical operators become autonomous agents, leveraging the brain's natural capacity for procedural memory and narrative retention over rote memorization.
bbstrader: Integrated SentimentAnalyzer NLP models into the flagship open-source trading toolkit for ML-driven signal generation, combining Python's analytical depth with C++ execution speed via a bidirectional pybind11 bridge.
Experience
A track record of technical leadership, from building AI infrastructure and founding engineering teams to driving $56M+ in trading volume worldwide.





Education
A multidisciplinary foundation spanning computer science, financial engineering, and quantitative methods, built across leading institutions worldwide.




Open Source
Production-grade tools for quantitative finance and high-performance market data processing, built in public, used by traders worldwide.
# C++ Core, Python Orchestrator via pybind11 from bbstrader.api import Mt5Handlers from my_strategies import MovingAverageStrategy # Inject Python MT5 handlers into C++ class strategy = MovingAverageStrategy(Mt5Handlers) if strategy.initialize(): while True: strategy.on_tick("EURUSD") # ML-driven sentiment signal from bbstrader.models import SentimentAnalyzer score = SentimentAnalyzer().analyze( "Fed hikes rates, markets soar!" ) if score > 0.7: print("Go long ๐")
[BACKTEST] bbstrader ยท MovingAverageStrategy โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ Period 2022-01-01 โ 2024-12-31 (36 mo) โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ Total Return +34.7% Benchmark +12.1% Sharpe Ratio 1.84 Sortino 2.31 Max Drawdown -8.2% Win Rate 61.4% Total Trades 847 Profitable 520 โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ [OK ] Pushing strategy โ MetaTrader 5 live [OK ] Connected ยท account #12345 ยท $50,000 [TICK] EURUSD=1.08412 signal=LONG lot=0.10 โ
import gzip from itchfeed import MessageParser # Filter: Add orders, Executions, Non-cross Trades parser = MessageParser(message_type=b"AEP") with gzip.open("01302019.NASDAQ_ITCH50.gz", "rb") as f: for msg in parser.parse_file(f): print(msg.decode()) # AddOrderNoMPIAttributionMessage( # stock=b'AAPL', price=1823400, shares=200, # buy_sell_indicator=b'B', timestamp=...) # [โ] 2,847,361 msgs ยท 3.42s
std::variant
for type-safe message dispatch, ideal for latency-sensitive HFT applications.#include <fstream> #include "itch/parser.hpp" itch::Parser parser; std::ifstream file("NASDAQ_ITCH50", std::ios::binary); // Streaming callback, zero-copy hot path parser.parse(file, [](const itch::Message& msg) { std::visit([](const auto& m) { std::cout << m << '\n'; }, msg); }); // Or collect filtered message types only file.seekg(0); auto msgs = parser.parse(file, {'A', 'E'}); // โ multi-GB/s ยท zero heap allocs in hot path
import sqlite3 conn = sqlite3.connect("economy.db") cur = conn.cursor() cur.execute(""" SELECT fm.time_period, fm.obs_value FROM financial_market fm JOIN countries c ON c.id = fm.country_id JOIN measures m ON m.id = fm.measure_id WHERE m.description = 'Short-term interest rates' AND c.code = 'USA' ORDER BY fm.time_period """) # ('2022-Q1', 0.08), ('2022-Q2', 1.21), # ('2022-Q3', 3.04), ('2023-Q1', 4.65) ...
include/ โ public headers ยท library contracts src/ โ implementation + CMakeLists tests/ โ Google Test ยท first-class citizens benchmarks/ โ Google Benchmark ยท perf analysis libs/ โ custom vcpkg overlay ports cmake/ โ build helper scripts # Configure & build (CMake Presets) cmake --preset mingw # or 'win' / 'cross' cmake --build build
// Observer ยท event-driven price feed struct PriceObserver { virtual void on_tick(double price) = 0; }; class MarketFeed { // Subject std::vector<PriceObserver*> subs_; public: void subscribe(PriceObserver& o) { subs_.push_back(&o); } void publish(double price) { for (auto* o : subs_) o->on_tick(price); } }; // โ RiskEngine, Logger, Strategy all notified // [23 patterns] Creational ยท Structural ยท Behavioral
Thought Leadership
Technical dispatches on distributed systems, fintech infrastructure, and algorithmic markets, published on TheFinTech Insights Digest.
Contact
Whether it's a custom trading system, an AI pipeline, a collaboration on open-source, or a conversation about markets and technology, reach out through any of these channels.