AI Engineer Founder Researcher
Hi, I'm Victor — nice to meet you.
I'm an AI engineer based in Nairobi, Kenya. I started in data science and analytics — forecasting, fraud detection, customer segmentation for fintech and financial services. When generative AI matured in 2022, I shifted to building production AI systems full-time: multi-agent platforms, RAG pipelines, on-device ML, multilingual assistants for clients across fintech, agriculture, logistics, and SMBs.
My work spans the full stack of applied AI — RAG chatbots that cite their sources, multi-agent pipelines that orchestrate dozens of specialized agents, forecasting engines powered by real statistical math (Holt-Winters, GARCH, OLS), on-device ML that runs on $80 phones, and multilingual assistants that work in Swahili and English. I take projects from zero to production — scoping, architecture, build, ship, and team training.
Five projects define my current work. Busara AI — a multi-agent data intelligence platform with a 7-stage analysis pipeline (v7.0, multi-agent system, PWA, live at busaraai.com). KilimoPRO — an offline-first app for 8 IGAD countries (300M+ people) with on-device crop disease detection, GARCH/VaR market forecasting, and a multi-LLM advisory council in Swahili and English. Verxlite — a universal AI workflow agent for sales & ops that automates follow-ups, CRM logging, and approvals across Gmail, Google Calendar, and HubSpot. Fraud Detection System — a real-time transaction fraud scoring engine in Go with a weighted-vote ensemble of statistical detectors, sub-millisecond latency, and full explainability for regulated finance. And AgentReplay — an open-source deterministic-replay debugger for AI agents that reproduces agent failures bit-exact with zero model calls. All five are real systems serving real users, not demos.
I also take on select client work — AI audits, custom agent systems, and full-stack builds. If you have a problem that needs real engineering (not a wrapper around ChatGPT), I'd love to hear from you.
End-to-End Builds
From scoping audit to production deployment. Full-stack: Next.js, FastAPI, Flutter, PostgreSQL, Docker, Kubernetes.
Multi-Agent & RAG
Multi-agent DAG orchestration, RAG with pgvector/Pinecone, hybrid search, reranking. LangChain, LangGraph, CrewAI.
Real Math, Not Vibes
Holt-Winters forecasting, GARCH/VaR, K-Means++, OLS regression, Granger causality — implemented from scratch.
Built for Africa
Offline-first, low-bandwidth, Swahili-capable. Systems that work on $80 phones and over SMS/USSD.
Academic foundation.
B.Sc. in Computer Science
Coursework in Data Structures & Algorithms, Machine Learning, Database Systems, Statistics, and Distributed Systems. Final-year project — an early crop advisory system that seeded the KilimoPRO concept. Top-quartile standing in the School of Computing.
Continuous Learning & Certifications
Active builder and maintainer of open-source AI projects (AgentReplay, KilimoPRO, IntelliFlow, Busara AI). Currently pursuing AI Engineer for Data Scientists Associate and AI Engineer for Developers Associate certifications — covering data management, model development, AI app deployment, and responsible AI. Aligned with Kenya's NASIP 2026-2030 agricultural strategy.
What I can build for you.
AI engineering for every stage — from a focused audit to a full multi-agent system. Built to ship, work in Swahili, and run on the devices your users actually have.
What I'm building right now.
A live snapshot of active projects, updated as I ship. No vanity metrics — just real work in progress.
Platforms I've built.
Five production systems I'm actively building and maintaining. Not demos — real users, real data, real consequences.
Busara AI
A multi-agent data intelligence PWA running a multi-agent system across a 7-stage parallel DAG pipeline (v7.0). Ingest → Engineer → Detect → Forecast → Infer → Report. Real statistical algorithms, not wrapper code. Live at busaraai.com with a Free → Pro → Team → Enterprise tier (99.9% SLA).
KilimoPRO
AI-powered agricultural intelligence for 8 IGAD countries (300M+ people). On-device crop disease detection (TFLite) for 15 crops, market forecasting with GARCH/VaR, multi-LLM advisory council in Swahili + English, 9 free keyless API endpoints, SMS webhook for feature phones. $0/month to operate — all 4 data sources free.
Verxlite
A universal AI workflow agent for sales & ops teams. Automates post-meeting followups, lead assignment, support triage, approval routing, and weekly summaries across Gmail, Google Calendar, and HubSpot. Multi-tenant SaaS with real auth, OAuth, encrypted tokens, Celery workers, and Langfuse observability. Shipped with a transparent 281-defect code review and full P0/P1 remediation in one week — 83 backend tests passing.
AgentReplay
Deterministic replay and counterfactual debugging for AI agents. Captures every LLM call, tool response, clock read, and RNG draw — then replays the run bit-exact with zero model calls. 0.67% overhead, 100% reproduction fidelity, 99 tests. LangGraph-first. Backed by a research whitepaper.
Fraud Detection System
A real-time transaction fraud scoring engine in Go — 7 statistical detectors (Z-Score, IQR, Velocity, Geo-Distance, Device Fingerprint, Merchant Risk, Behavioral Anomaly) fused by a weighted-vote ensemble, calibrated by logistic regression, and overlaid with a deterministic rules engine. Scores in ~33 µs per transaction. Full case-management workflow, RBAC auth, Prometheus metrics, OTel tracing, and Slack alerts. 84% recall, 1.37% false-positive rate on 1,000 labelled transactions. Built for PCI-DSS / PSD2 / GDPR Article 22 compliance.
IntelliFlow (Busara AI)
33-agent self-evolving multi-agent data intelligence platform. 7-stage pipeline with trajectory-based learning inspired by the AReaL paper. Turborepo monorepo, Next.js 15.
GitHub →Financial Forecasting Suite
Production-grade time series forecasting toolkit. Holt-Winters, ARIMA/SARIMA, Prophet-style decomposition, ensemble forecasting with auto model selection. 87.3% mean accuracy across 12 financial metrics.
GitHub →Fraud Detection System
Go microservice for real-time transaction fraud scoring. Z-Score + IQR + Velocity detectors in a weighted-vote ensemble. Scores transactions in ~50–300µs. Explainable alerts for regulated finance.
GitHub →Customer Segmentation Engine
RFM analysis + K-Means + Gaussian Mixture Models for actionable customer segmentation. 6 optimal clusters, 89% month-over-month stability, 23% lift in campaign response rate. Built for financial services.
GitHub →Tools I build with.
Click a category to filter. Every tool here is something I've shipped to production — from multi-agent DAGs to on-device ML.
Recent GitHub activity.
Live from my GitHub — what I've been pushing lately. Updated in real-time via the GitHub API.
Work that travels.
Public repositories, real commits, verifiable impact. No vanity metrics — just code shipped and research published.
AgentReplay — deterministic replay for AI agents. 99 tests, 0.67% overhead, 100% reproduction fidelity. MIT-licensed, whitepaper published.
KilimoPRO — offline-first AI for 8 IGAD countries (300M+ people). On-device crop disease detection for 15 crops, GARCH/VaR market forecasting, multi-LLM advisory council. $0/month to operate.
Verxlite — universal AI workflow agent for sales & ops. FastAPI + Celery + Next.js, Google Workspace & HubSpot OAuth, encrypted tokens, Langfuse observability. 83 tests passing after a transparent 281-defect code review.
Fraud Detection System — real-time transaction scoring in Go. 7 statistical detectors, weighted ensemble, logistic calibration, rules engine, case management, RBAC, Prometheus, OTel. 84% recall, 1.37% FPR, ~33 µs per score. PCI-DSS / PSD2 / GDPR Article 22 ready.
Busara AI — multi-agent data intelligence platform. Multi-agent system, 7-stage DAG pipeline, PWA, 4 pricing tiers. Live at busaraai.com serving real users.
10 long-form technical articles on multi-agent AI, agricultural intelligence, statistical engineering, and agent reliability research. 22,000+ words of documented work.
Let's build something.
Open to collaborations on AI products, agricultural technology, and ambitious African tech.
Book a 30-minute call
Not sure where AI fits? Let's talk. We'll discuss your goals, audit where AI can help, and recommend the right next step — even if that's "do nothing yet." No pressure, no sales pitch.
Prefer email? Email me directly and we'll find a time.