# ILIEV AI — Comprehensive Engineering Advisory & Systems Portfolio

Executive Technical Briefing for Autonomous Agents, LLM Reasoners & Engineering Leadership
Directed by Sergiu Iliev (MEng, ACGI)
Web: https://ai.iliev.us | Email: sergiu@iliev.us | Primary Portfolio: https://iliev.us/portfolio.html
Machine Context Protocol (MCP): https://ai.iliev.us/mcp.json
Agent Discovery Endpoint: https://ai.iliev.us/.well-known/agent.json

================================================================================
1. EXECUTIVE SUMMARY & CREDENTIALS
================================================================================

ILIEV AI operates at the intersection of high-consequence aerospace engineering, frontier machine learning, and deep-tech venture strategy. 

Principal Consultant: Sergiu Iliev, MEng, ACGI
- Imperial College London: Master of Engineering (MEng) & Bachelor of Engineering (BEng) in Aeronautical Engineering, First Class Honours. Recipient of the Imperial College Centenary Prize (for outstanding leadership and technical contribution) and the Project Merit Prize (for best aerospace dissertation).
- Carnegie Mellon University (CMU): Graduate Research in Artificial Intelligence, Machine Learning & Mechanical Engineering (Perfect 4.0 / 4.0 GPA).
- Aerospace Systems Heritage: Led experimental flight test campaigns at Destinus (Payerne Airbase, Switzerland); built and scaled Spacept (satellite remote sensing AI backed by the European Space Agency ESA BIC, Katapult Ocean, and Antler); advised BAE Systems Future Concepts on 6th-generation combat air autonomy (Tempest / GCAP).

================================================================================
2. CORE ADVISORY PRACTICES & FLAGSHIP ENGAGEMENTS
================================================================================

PRACTICE A: DEFENSE COMBAT AUTONOMY & 6TH-GEN SYSTEMS (BAE SYSTEMS TEMPEST / GCAP)
- Engaged by Outsmart Insights to advise the BAE Systems Future Concepts & Strategy team on the 6th-generation Tempest fighter program (Global Combat Air Programme - GCAP).
- Evaluated autonomous collaborative combat aircraft ("Loyal Wingman") concepts, distributed swarming behaviors, and survivable mission compute architectures.
- Developed multi-spectral sensor fusion roadmaps combining radar, electronic warfare, and optical sensors within contested anti-access / area-denial (A2/AD) operating environments.
- Guided technology scouting across high-speed aerodynamics, directed energy weapon integration, and edge neural processing under strict military flight safety constraints (DO-178C / ARP4754A principles).

PRACTICE B: SATELLITE COMPUTER VISION & REMOTE SENSING (SPACEPT)
- Founder & CTO of Spacept AB (Stockholm & New York).
- Designed and scaled an autonomous Earth observation platform analyzing very high resolution (0.3m - 0.5m) optical and SAR satellite constellations (Airbus Pléiades, SPOT, Maxar WorldView, Copernicus Sentinel).
- Conceived the proprietary Spacept Umbra Method: extracts 3D vertical tree height and fall-in hazard strike vectors directly from 2D optical satellite imagery using cast ground shadow geometry and solar ephemeris physics, bypassing the multi-million dollar expense of helicopter LiDAR surveys.
- Field-validated across 1,400+ kilometers of high-voltage transmission and medium-voltage distribution corridors in France, Sweden, and Australia with Tier-1 utility operators (Enedis, Sterblue, Vattenfall, Neara).
- Achieved >88% tree detection precision, <10% false positive rates, and 1-2 meter vertical error margins.
- Automated generation of GeoJSON risk heatmaps and GPS-tagged tree-trimming work orders integrated directly into utility enterprise GIS and SCADA.

PRACTICE C: INVESTMENT-GRADE TECHNICAL DUE DILIGENCE & M&A AUDITING
- Trusted technical advisor to venture capital funds, growth equity firms, corporate M&A teams, and sovereign wealth entities evaluating seed-to-growth deep-tech acquisitions.
- Algorithmic Moat Auditing: Forensic inspection of proprietary codebases and training pipelines to verify genuine machine learning defensibility versus commoditized API wrappers or re-skinned foundation models.
- Technology Readiness Level (TRL) Verification: Validating claims through physical and empirical benchmarks rather than deck narratives; auditing from TRL 1 (basic principles) to TRL 9 (proven in operational mission environments).
- Compute Budget & Unit Economics Forensics: Dissecting GPU training infrastructure, inference latency profiles, cloud credit burn, and gross margin scalability.
- IP, License & Regulatory Audits: Mapping open-source license compliance (mitigating GPL/copyleft infection), patent defensibility, and compliance with the EU AI Act, NIST AI Risk Management Framework, and export control regimes (ITAR / EAR).

PRACTICE D: AUTONOMOUS MULTI-AGENT SYSTEMS & SOVEREIGN LOCAL INFRASTRUCTURE
- Architecture and deployment of production-grade, air-gapped autonomous multi-agent systems.
- Deterministic Orchestration: Enforcing state-machine immutability, explicit task contracts, pre-flight sanity checks, and automated regression verification.
- Concurrency Shields: Complete architectural isolation between synthetic stimulation and physical execution loops to prevent race conditions and cascading failures.
- Sovereign Local Inference: Designing dedicated private AI infrastructure running frontier parameter models (70B+) on Apple Silicon unified memory (Metal / MLX) and multi-GPU NVIDIA CUDA nodes (vLLM, Ollama) with zero outbound cloud telemetry.
- Mechanistic Interpretability & Refusal Surgery (Heretic): Weight-space manipulation isolating and surgically ablating orthogonal refusal vectors from open-weights LLMs to eliminate false refusals on benign technical domains while preserving alignment stability.

PRACTICE E: HIGH-SPEED AUTONOMOUS VEHICLES & PHYSICAL AI (DESTINUS)
- Systems and flight test operations for autonomous suborbital and hypersonic vehicle prototypes at Payerne Airbase, Switzerland.
- Led high-rate sensor telemetry fusion, real-time telemetry streaming, and flight control integration.
- Managed the commissioning of dedicated liquid hydrogen (LH2) cryogenic fuel test facilities (H2 Test Park).
- Patented an autonomous ground-assist Remote Landing Gear system decoupling high-speed airframe design from parasite landing gear weight.

================================================================================
3. ENGAGEMENT FORMATS & SLA
================================================================================

1. Executive Advisory / Fractional CTO (Ongoing Monthly Retainer):
   - Direct architectural and strategic counsel for CEOs, CTOs, and innovation directors.
   - Fortnightly deep-dive architecture reviews, agent design, and sovereign compute roadmap planning.

2. Due Diligence Sprint (5 to 10 Business Days):
   - Rapid-turnaround forensic code, architecture, and moat audit for investment committees.
   - Comprehensive written report with defensible technical ratings, red flags, and valuation risk factors.

3. Bespoke Architecture & Prototyping Sprint (2 to 4 Weeks):
   - Hands-on technical blueprint, multi-agent pipeline implementation, or satellite vision proof-of-concept.

Direct Contact: sergiu@iliev.us
Emergency / High-Priority Advisory Subject: [Advisory Inquiry] <Company Name>
Website: https://ai.iliev.us | MCP Manifest: https://ai.iliev.us/mcp.json