Nedu Ekeruche

Toronto, ON

Nedu Ekeruche

Software Engineer | Quantitative Developer | FinTech

I build software systems at the intersection of backend engineering, financial technology, and applied machine learning.

With a background in Mathematics and Statistics and experience developing production backend services, I enjoy solving problems involving distributed systems, financial modelling, data-intensive applications, and intelligent analytics.

My work includes real-time analytics infrastructure, reinforcement-learning-based portfolio optimization, and graph neural network approaches to financial transaction monitoring.

Core interests: Backend Engineering · Quantitative Development · Financial Technology · Machine Learning

Projects

Deep Reinforcement Learning for Portfolio Allocation

Reproducing J.P. Morgan AI research to compare PPO against classical mean-variance optimization under market volatility.

  • Reproduced a J.P. Morgan AI research setup to test whether a learning-based portfolio manager can allocate across market sectors as well as a classical optimizer.
  • On ten held-out years of S&P 500 sector data (2012–2021), the agent matched the optimizer on Sharpe ratio (0.96 vs 0.94) and cut the worst drawdown (−15.4% vs −36.3%).
  • Built a Gymnasium market-replay environment over those sector series, with softmax long-only weights and a Differential Sharpe reward so the agent is scored on risk-adjusted return.
  • Trained a Stable-Baselines3 PPO policy and benchmarked each held-out year against a Ledoit–Wolf max-Sharpe mean-variance baseline in PyPortfolioOpt.
  • Shipped a FastAPI backtesting API and a Vite/React dashboard for equity curves, rolling Sharpe, and allocation heatmaps.
  • Python
  • PyTorch
  • Stable-Baselines3
  • Gymnasium
  • PyPortfolioOpt
  • FastAPI
  • React
  • Vite

Graph Neural Network AML Transaction Monitoring

Real-time financial crime intelligence: HeteroSAGE scoring over synthetic and IBM AML transaction graphs.

  • A live monitor that scores transactions for money-laundering risk and shows an investigator how funds moved between accounts.
  • Runs on generated test traffic and on the public IBM anti-money-laundering dataset, so the same screen can simulate a feed or replay a known case.
  • Modelled banks, crypto venues, and cross-border entities as a heterogeneous graph and scored flows with HeteroSAGE in PyTorch Geometric.
  • Built a synthetic multi-network generator that injects known laundering typologies, and converted IBM AML CSV exports into replayable JSONL event streams.
  • Streamed live scores over WebSockets to a Vite/React dashboard with graph visualization, alerts, and a switch between the simulator and dataset replay.
  • Python
  • PyTorch Geometric
  • HeteroSAGE
  • WebSockets
  • React
  • Vite
Other ProjectsShow

Craftha

Marketplace for verified handmade crafts, with separate paths for makers and collectors.

  • A marketplace for verified handmade work, closed to dropshipping and AI-generated goods.
  • Built the Next.js site with separate paths for makers and collectors, and an early-access waitlist ahead of the public marketplace.
  • Next.js
  • TypeScript

SereinMod

Developer API for filtering toxic comments and media instead of reviewing them by hand.

  • Lets a product screen comments and uploads for toxicity instead of reviewing each item by hand.
  • Built the site and a REST API that scores text and media by category and returns an allow-or-deny decision.
  • Documented the integration with request examples and tiered pricing so a developer can call it with an API key.
  • Next.js
  • TypeScript

Whot Online

Real-time multiplayer Whot, with private rooms, spectators, and a trained computer opponent.

  • Play Whot online with friends in a private room, watch as a spectator, or play against the computer.
  • Built a Next.js table with drag-and-drop cards, room create and join, and live match state over WebSockets.
  • Kept the rules and match state in Redis behind FastAPI, including turn timers, spectators, and specials such as Pick Two, Hold On, and General Market.
  • Trained a Gymnasium self-play policy with Stable-Baselines3 and served it as the computer opponent on that same rules engine.
  • Next.js
  • TypeScript
  • FastAPI
  • Redis
  • WebSockets
  • PyTorch
  • Stable-Baselines3
  • Gymnasium

Experience

  1. Jan 2024 – Sept 2026

    Software Developer

    SevenDot Analytics

    Python, FastAPI, Redis, WebSockets, PostgreSQL

    • Built the streaming backend behind production analytics, so live results reach users as new events arrive.
    • Designed FastAPI microservices and low-latency WebSocket streams for those analytics workloads.
    • Used Redis for short-lived, event-driven state and PostgreSQL for durable records.
  2. Since June 2025

    Freelance Lead Developer

    ThriveParth

    Website

    Next.js, TypeScript, Cloudflare R2, Resend

    • Built the public site for a Toronto nonprofit supporting newcomers and youth: programs, events, volunteer and support intake, and a community gallery.
    • Shipped an admin for one-time, weekly, and monthly events and for gallery folders. Image uploads go to Cloudflare R2 through presigned URLs, so the browser writes to storage directly.
    • Routed contact, support, volunteer, and newsletter forms to the team through Resend.

    Shutter Unit

    Website

    Next.js, TypeScript, Cloudinary

    • Built a photography site for weddings, portraits, and commissions, with collection pages, a film feature, and a contact path.
    • Served the galleries from Cloudinary so each frame is resized and compressed on the CDN instead of shipping full-resolution files.

Education

  1. Expected Graduation April 2027

    Financial Technology

    Sheridan College

  2. Graduated June 2023

    B.Math & Statistics

    Carleton University

Credentials: Google Professional Machine Learning Engineer · Google TensorFlow Developer

Skills

Systems & Backend

FastAPI · WebSockets · Redis · PostgreSQL · Celery · Docker · CI/CD

Quantitative & Financial

NumPy · Pandas · PyPortfolioOpt · Excel

Machine Learning

PyTorch · PyTorch Geometric · Stable-Baselines3 · Gymnasium

Languages

Python · C# · TypeScript · Go · SQL

Frontend

React · Next.js · Vite · TypeScript · Tailwind CSS