Aileron Technologies

Cyber risk is business risk. We turn it into finished engineering work.

Regulators hold the management body accountable (NIS2), the EU AI Act fines reach 7% of turnover, and your first enterprise customer sends a security questionnaire long before either. We close the gaps in code and infrastructure and hand over the evidence — not slides.

20+ years experienceFormer CTO & Chief ArchitectAnthropic Cyber Verification Program: Defense AccessISO 27001 & ISO/IEC 42001 technical readiness
Security questionnaire blocking a deal?Bring it to a 30-minute call — you leave knowing what is a day of work and what is a month
New: Agent Control. A fixed-price audit of what your coding agents wrote.
01

Founder

Aleksei Kudriashov

Software Architect, CTO & Founder

Software architect, CTO and founder with more than 20 years designing, securing and delivering high-performance software systems. I work directly with founders, investors and engineering leaders on the decisions where mistakes are expensive — security and compliance before a deal, architecture, scale, and delivery under real constraints.

Background
  • Telecom platforms
  • SaaS products
  • Aerospace systems
  • AI platforms
  • Security & compliance engineering
02

Selected Experience

Telecom Platforms

Chief Architect for distributed telecom systems processing over 1 million calls per week, with availability and latency held to strict targets.

SaaS Products

Built and led SaaS platforms to scale — including a taxi-services platform supporting 1.5 million orders per month and 7 000 drivers online.

Aerospace Systems

Co-founder of an aerospace software company building cloud-native platforms for flight data and fleet intelligence; built a zero-trust, certificate-based device-connectivity system for in-flight data streaming.

AI Platforms

Directed R&D and architecture for AI products, retrieval systems and agent platforms, from data pipelines to production serving — including the controls auditors now ask about: data masking, prompt versioning, access inside RAG.

04

Engagement Models

Fixed-scope engagements designed to close a security gap, reduce technical risk and accelerate decision-making.

How engagements work
Most Requested

Architecture Risk Audit

Independent assessment of software architecture, scalability risks, technical debt and operational bottlenecks.

Ideal for
SaaS foundersCTOsEngineering leaders
Deliverables
  • Architecture review
  • Risk assessment
  • Technical debt analysis
  • Prioritized recommendations
Typical duration1–2 weeks
Book an Architecture Review
Start Here

Compliance Gap Check

Three days on your customer's questionnaire or target standard, your code and your infrastructure, with our open-source architecture-aware code analysis (aileron-tech/wingbox). You leave with a prioritised plan: what is a day of work, what is a week, what is a month.

Ideal for
B2B SaaS foundersAI startupsCTOs
Deliverables
  • Requirement-to-gap mapping
  • Prioritised remediation plan
  • Effort estimate per gap
  • Evidence checklist
Typical duration3 days
Book a 30-Minute Review
Pre-Deal

Compliance Remediation Sprint

We close the technical controls behind SOC 2, ISO 27001 and ISO/IEC 42001: secrets, access and MFA, logging, backups, encryption, vulnerability handling, CI/SDLC evidence and LLM data flows. Evidence goes into your compliance platform (Vanta, Drata, Secureframe or any other) or into a package for the auditor.

Ideal for
B2B SaaSAI productsTeams facing a questionnaire
Deliverables
  • Closed technical controls
  • Evidence in your tool or as a package
  • LLM data-flow controls: PII masking, prompt versioning, RAG access, fallback
  • Handover notes
Typical duration10 days over 2–3 weeks
Discuss a Remediation Sprint
Investor Service

Technical Due Diligence

Independent evaluation of software assets, architecture quality, delivery capability and technical risk.

Ideal for
InvestorsAcquirersBoardsStartup founders
Deliverables
  • Executive summary
  • Technical findings
  • Risk matrix
  • Recommendations
Typical duration2–3 weeks
Discuss Due Diligence
High Impact

Recovery Sprint

Rapid assessment and recovery planning for delayed, over-budget or technically troubled software initiatives.

Ideal for
FoundersCTOsProduct leaders
Deliverables
  • Project assessment
  • Delivery risk analysis
  • Recovery roadmap
  • Leadership recommendations
Typical duration2–4 weeks
Discuss Recovery Options
Ongoing

Compliance Retainer

Two days a month to keep controls in shape between audits, answer customer questionnaires and auditor questions, and review changes before they open new gaps.

Ideal for
Teams after a first auditStartups selling to enterprise
Deliverables
  • Monthly control review
  • Questionnaire answers
  • Auditor Q&A support
  • Change review
Typical duration2 days per month
Discuss a Retainer
Experience Behind These Engagements
  • 20+ years building and securing software systems
  • Former CTO and Chief Architect
  • ISO 27001 and ISO/IEC 42001 technical readiness for AI and SaaS products
  • Zero-trust PKI and ground services for in-flight data streaming at a US avionics manufacturer
  • Telecom platforms handling over 1 million calls per week
  • SaaS platforms supporting more than 1.5 million orders per month
  • AI and R&D leadership across Europe and North America
  • Inventor on a French patent application (INPI, 2026)
05

Case Studies

Enterprise AI

AI-First Workspace with GDPR-Compliant RAG

ContextA Marseille AI company building an AI-first workspace: AI operatives grounded in the company’s own mail, calendar, channels, tasks and files. Role: Director of R&D, February 2025 to October 2026, team grown from 7 to 15.

Challenge

Ship agents, retrieval and automation as one coherent product, on company data, with auditable actions and GDPR, ISO 27001 and SOC 2 expectations from enterprise prospects.

Approach

Defined and steered the R&D strategy: a scalable GDPR-compliant RAG on containerized Python services, a mutable vector index with proximity-aware compaction and LSH-Bloom routing, hybrid lexical-vector search with structured chunking, multi-agent orchestration, and a multi-tenant MCP execution sandbox (containers, RBAC, network policies, seccomp, tracing).

Outcome

Median search time cut by more than half with recall above 95 %; orchestration measured above 99.5 % success with sub-3-second time to first token under load. Prix Data & IA (C’est la Tech BB, BFM Business) and Prix de la Transformation Numérique (SEPEM) in June 2025.

ReferenceDirector of R&D of the company (2025–2026).
Avionics · Security

Zero-Trust PKI and Ground Services for In-Flight Data Streaming

ContextA US manufacturer of aircraft data collection and transmission equipment (as contractor, through an aerospace software company). In February 2026 the client publicly demonstrated real-time streaming of aircraft sensor data to ground analytics during flight.

Challenge

Authenticate every onboard unit without a VPN, isolate each device to its own data, and ingest, store and forward the streams with ordering and session boundaries preserved — at mission-critical reliability.

Approach

Built a certificate-based PKI: factory certificate on every unit, production certificate obtained over mutual TLS, broker-side authorization deriving identity and scope from the certificate, per-device topic isolation and revocation. Delivered the ground services for ingestion, storage and forwarding, a secure file-transfer path, benchmark and load-test harnesses, and the production rollout of the PKI.

Outcome

A maintainable, compliant platform on Kubernetes, MQTT over TLS 1.3 with client certificates, TypeScript and .NET, specification-first with requirement traceability.

ReferenceSecurity and ground-side architect, avionics OEM programme (client under NDA).
Aerospace

Open-Source Flight Control Stack and MBSE Toolchain

ContextA US aerospace software company: co-founder and board member since January 2021, leading the software side of the company.

Challenge

Deliver certifiable flight software for unmanned aerial vehicles across changing hardware, with a toolchain small teams can actually operate under ARP4754A and DO-178C.

Approach

An autopilot stack in C, hardware- and OS-agnostic, extended through atomic functions generated from formal descriptions and configured per vehicle through XML domain-specific languages (control laws, guidance state machines, trajectory patterns, navigation state estimation); a flight-computer product line (RS-485, CAN, ARINC-429, MIL-STD-1553B); a collaborative online platform with MBSE services and report generation from the formal configuration.

Outcome

Published as open source under the BSD-3 licence and integrated with partner avionics.

ReferenceCo-Founder / VP SaaS of the company.
Aerospace R&D

Self-Supervised Engine Health Monitoring (Patent Application FR2604268)

ContextAileron Technologies R&D, February to April 2026. French patent application FR2604268 filed at INPI on 2 April 2026, inventor and applicant Aleksei Kudriashov.

Challenge

Supervised remaining-useful-life methods need failure labels, proprietary models need retraining per engine variant, and foundation models need cloud compute. Predictive maintenance had to run embedded and apply zero-shot to new operating conditions.

Approach

A minimal encoder-decoder transformer with two-dimensional attention heads, trained by causal self-supervision on raw multivariate sensor windows, with a physics-informed positional encoding on operating cycles; a frozen encoder serves forecasts, health and deviation scores, novelty detection, uncertainty bands and a remaining-useful-life head.

Outcome

About 131 000 parameters, under 1 MB of weights, about 32 ms inference. On the NASA C-MAPSS benchmark: forecasting RMSE 2.29 at horizon 10 versus 2.63 for TimesFM; RUL RMSE 12.1 on FD003 and 14.1 on FD001.

ReferenceInventor, Aileron Technologies — research, training, evaluation, dashboard and patent text built end to end.
Legal AI

Multi-Agent RAG for EU Regulatory Analysis

ContextA Paris legal-technology platform analyzing European regulatory requirements for compliance teams. Role: AI Lead on contract, October 2024 to April 2025.

Challenge

Manual legal analysis and document preparation created significant overhead, with complex regulatory texts requiring expert interpretation at scale.

Approach

Designed and implemented a multi-agent, retrieval-augmented architecture that ingests EU directives and regulations, extracts structured legal obligations through staged analysis and supports AI-assisted drafting; contributed to the technical pitches that secured key partnerships.

Outcome

Reduced the effort required for document preparation and enabled scalable, structured analysis of regulatory material.

ReferenceAI Lead (contract).
Telecom

Call-Queue Platform Behind the Sochi 2014 Olympic Hotline

ContextA contact-centre outsourcer with more than 1 500 operators and about one million calls a week, won the tender for the official multilingual support hotline of the Sochi 2014 Winter Olympic Games.

Challenge

Carry the Olympic traffic with high availability, predictable latency and real-time, skill-based routing under sustained load: the hotline ran from 1 November 2013 to 31 March 2014 and handled more than 100 000 calls in seven languages.

Approach

As Chief Architect (2013–2015) designed and led the call-queue platform: an actor-based distributed call-distribution architecture with skill-based routing, priorities and real-time queue management, Asterisk sharding, resilient integration with the telephony layer and a clean control API for operator desktops, supervisors, wallboards and reporting.

Outcome

Over one million calls per week with availability and latency held to strict targets; the hotline earned the top “Crystal Headset” award in the Best Outsourcing Partner category.

ReferenceChief Architect of the outsourcer.
Mobility SaaS

Taxi-Services SaaS at 1.5 Million Orders per Month

ContextA taxi-services SaaS platform serving dispatch companies and their drivers, architected in the same period as the call-queue platform (2013–2015).

Challenge

Rapid growth across cities, companies and devices required an architecture that could scale order volume while staying reliable in real time, with multi-tenant isolation.

Approach

Designed a microservices architecture with polyglot persistence, event-driven processing and real-time updates, device connectivity for drivers and multi-tenant isolation; code analysis, profiling and performance tuning of the legacy telecom software around it.

Outcome

Supported 1.5 million orders per month and 7 000 drivers online while maintaining operational stability.

ReferenceSoftware Architect of the platform.
06

Credibility

Aileron is not a software development agency and not a paper auditor. It is an independent engineering practice: we close the security and compliance gaps in code and infrastructure, and we help founders, investors and engineering leaders make high-stakes technical decisions with confidence.

  • 20+ years building and securing software systems
  • Former CTO and Chief Architect
  • Architected telecom systems handling over 1 million calls per week
  • Led SaaS platforms supporting over 1.5 million orders per month
  • Built a zero-trust PKI and ground services for in-flight data streaming at a US avionics manufacturer
  • Directed R&D for AI-driven products and retrieval systems
  • Inventor on a French patent application (INPI FR2604268, 2026) for self-supervised predictive health monitoring
  • Open-source contributor to Archestra, an enterprise AI platform (MCP gateway, registry and guardrails): built its Browser MCP feature, merged in January 2026
  • Data & AI prize at BFM Business “C’est la Tech” 2025 and second place at one of the world’s largest AI hackathons
  • Approved organization in Anthropic’s Cyber Verification Program (Defense Access, 2026): frontier models with cyber safeguards relaxed for verified defenders
  • Author of aileron-tech/wingbox (MIT): open-source architecture-aware code search, dependency graphs and impact analysis for fourteen languages and frameworks — the engine behind our code-level audits
07

Contact

Let’s talk about the decision in front of you — a security questionnaire blocking a deal, an architecture review, a due diligence, or a project that needs to get back on track. Bring the questionnaire: in 30 minutes you will know what is a day of work and what is a month.

Pick your situation — the message writes itself, edit what you like:

Paste a Drive, Dropbox or Notion link, or email the files to contact@aileron-tech.fr after sending this form.

Prefer to talk? Pick a 30-minute slot

Calls are held in English only.

Book a 30-Minute Review