Fractional CTO — Startups · FinTech · Quant AI
Hossayne Tahiri
I help early-stage founders ship complex products without hiring a full-time CTO. 15 years translating business goals into shipped products and leading the technical teams that deliver — inside tier-1 banks (Moody's, BNP Paribas, Société Générale), as CTO of Everdiams (RWA tokenization), and as CEO of Mandeo (PropTech marketplace).
~$SHELL 15_years.of(quant_finance) + 7_years.of(building)
Available for new engagements · Starting July 2026
01 —
What I do
Strategic partnership — roadmap to board
Turning a founder's vision into a technical roadmap. Arbitrating between speed and scalability. Being the technical voice in board meetings, investor pitches, and tech due diligence.
STRATEGY · PRODUCT · BOARD
Team & engineering leadership
Hiring your first engineers, arbitrating technical choices against business objectives, and keeping deliveries on time. The senior layer your team needs, without the full-time cost.
LEADERSHIP · HIRING · DELIVERY
Quantitative & technical depth
When the problem is deeply technical — high-performance simulation engines, pricing models, risk systems, cloud architecture — I go hands-on. 15 years shipping production-grade systems inside tier-1 banks.
ARCHITECTURE · QUANT · CLOUD
02 —
How I operate
Fifteen years of shipping inside banks, startups, and regulated crypto. Five principles I keep coming back to.
Breadth is leverage.
I've shipped balance-sheet engines in Python, full-stack marketplaces in Next.js, and cloud workloads on AWS. That range lets me challenge, stress-test, arbitrate — and match the right competence to the right problem. Specialists answer inside their frame. I work across frames.
Translate, don't just ship.
On Mandeo I owned the translation from business and strategic objectives into an actual product — and kept that coherence end-to-end: roadmap, architecture, pricing engine, UX, go-to-market. A fractional CTO who can't hold that bridge ships fast in the wrong direction.
Decide at 70% information.
Early-stage startups never have the data to decide cleanly. I make the call at 70%, document the reasoning, and move. Correctness beats perfection when runway is burning.
Full operating range — code or direct.
I shipped PASS-ALM as primary quant lead developer inside a tier-1 bank. I ran Mandeo end-to-end without writing the last line of code myself — product, platform, go-to-market. A fractional CTO who can only operate one mode is half the value.
Diplomas, successes, and scars.
École Polytechnique, Sciences Po and Télécom Paris gave me the frame. Fifteen years in the field — inside banks, regulators, and startups — gave me the rest: the pattern-matching on what works, and on what breaks. Architectures that collapse at scale, teams that implode, launches that slip: I've been through them. You hire me for all three.
03 —
Selected work
Real Estate PropTech — Mandeo
2025 →2025 — present · Co-founder & CEO
A French marketplace dedicated to empowering home sellers by enabling the sale of their listing to real estate agents. Co-founder and CEO — full ownership of strategy, product, team, and delivery. What the role covers: — Translating business and strategic objectives into a coherent product, roadmap, and go-to-market plan — Product decisions, UX, branding, pricing strategy, commercial positioning — Directing the technical build end-to-end and aligning contributors (engineering, design, legal, marketing) around shared priorities — Holding coherence between strategy, product, and technical execution — from day one to live operations Technical delivery under that direction: full-stack marketplace (authentication, listings, bidding engine, agent dashboards, admin back-office), proprietary pricing engine combining statistical modeling on French DVF data with machine learning (Kernel Density Estimation + XGBoost), AWS infrastructure with CI/CD, monitoring and analytics.
Python · FastAPI · Next.js · PostgreSQL · AWS · XGBoost · KDE
Group ALM Simulation Platform — Top-Tier European Bank
2020 →2020 — present · Via Krypteros
End-to-end design and deployment of a proprietary ALM platform replacing the legacy vendor (Moody's Analytics). Now used daily by 100+ ALM analysts across both retail networks of a top-tier European banking group. I led the build of three distinct layers as lead architect and primary developer: — PASS-ALM: the core quantitative engine (Python / Polars / NumPy). Full balance sheet modeling (interest rate products, non-maturing deposits, regulated savings, caps/floors/swaptions), pricing models (Bachelier, Shifted Black), FTP, and key indicators (NII, EVE, GAP, sensitivities). 10× faster computation and tenfold simulation capacity vs. the legacy vendor. — Cloud infrastructure: GCP-native architecture for heavy simulation workloads. Benchmarked and designed the execution stack across Cloud Run, Cloud Batch, and GKE Standard. Massive cost reduction at scale while enabling scenario analysis at a granularity previously out of reach. — kALM: a proprietary desktop application (React / Tauri / Rust) giving ALM analysts a modern UI to run scenarios, edit assumptions, and visualize outputs. Versioned releases, native performance, offline-capable, distributed across the group's retail networks. Technical lead and cross-silo coordinator on a multi-year, group-scale build. Directed a 5-to-15 person cross-functional team (quants, developers, ALM experts) — scoping, hiring, architecture calls, code reviews, production rollout, and change management — while keeping ALM business, IT, risk, and executive stakeholders aligned from roadmap to rollout. Development heavily augmented with AI (Claude, Claude Code, agentic workflows) — a core part of how I ship code, models, and documentation at senior-engineer pace.
Python · Polars · NumPy · GCP (Cloud Run · Cloud Batch · GKE) · React · Tauri · Rust · Claude Code
RWA Tokenization of Physical Diamonds — Everdiams
2019 → 20212019 — 2021 · Founder & CTO
One of the earliest Real-World Asset tokenization projects in Europe. As CTO, I led the full technical organization across several teams: — Data science team with Institut Louis Bachelier — proprietary ML pricing models (neural networks, XGBoost) for real-time valuation of diamond-backed tokens — Front-end and back-end engineering teams — public platform and investor dashboard — Smart-contract partnership with Talium — ERC-20 tokens on Ethereum and the on-chain issuance flow — Regulatory track at the French AMF — full PSAN/DASP dossier, anticipating the EU MiCA framework
Solidity · Ethereum · Python · TensorFlow · XGBoost · Next.js
DeFi Risk Models — Unslashed Finance
2020 → 20212020 — 2021 · Senior Consultant via Krypteros
Designed and calibrated de-peg risk models for stablecoins (USDT, USDC, WBTC) using ARMA and stochastic diffusion. Built a scoring framework for the probability of exploits on governance tokens, plus an elasticity model for price impact of a hack.
Python · Stochastic Processes · DeFi
04 —
Background
I started my career at the French Treasury, analyzing the macroeconomic impact of tax reforms. From there, I moved to Moody's sovereign rating desk in London, then to tier-1 European banks — BNP Paribas, Société Générale, and currently a long-term engagement inside a major European banking group, where I lead the design of a group-wide ALM platform.
In parallel, I've founded and led two products: Everdiams (RWA tokenization of physical diamonds, 2019–2021) as CTO directing engineering, data-science and smart-contract teams, and Mandeo (PropTech marketplace, 2025) as co-founder and CEO, owning strategy, product and delivery end-to-end.
05 —
Now
Status — available
Currently leading Mandeo as CEO — the PropTech marketplace I co-founded.
Still with bandwidth for additional fractional CTO engagements, 20–30 hours per week, starting immediately. Full availability from July 2026.
06 —
Let's talk
If you're a founder or a company building an ambitious, complex project — and you want a CTO-level partner without the full-time cost — I'd love to hear about it.