Most of what I build is software. The center of gravity is software — systems that take chaos, open-ended goal and carry it all the way to a finished engineering: translating entire books while keeping their formatting intact, dubbing and localizing video courses across languages, running multi-source research, drafting long documents, and automating the parts of knowledge work that never scaled by hand.

Around those agents sits the rest of the software — the AI systems, the content and media pipelines, the publishing and learning platforms, the sales engine. Different surfaces, same underlying work: understand the problem, design the architecture, and write the code that makes it run reliably.

• Babel — Advanced document translator that localizes complex technical documents across languages while preserving structure, formatting, and meaning.
• Octavia — Advanced video translator and multilingual course localization AI — a Multi-Modal RAG system for end-to-end video translation and dubbing, integrating Azure, AWS, and OpenAI (Whisper / TTS) models for high-quality multilingual output.

LUNARTECH LABS is the applied-AI product studio behind LunarTech's software portfolio — translation engines, knowledge systems, media platforms, and agentic developer tooling. Each product below is a standalone system built and shipped by the LABS team.

• Aura — LunarTech's book-publishing arm, producing and releasing full-length books and handbooks.
• Babel — Document and book translation tool combining Agentic OCR with LLMs to translate while preserving original layout, tables, and formatting.
• Babel Extreme — End-to-end pipeline that translates scanned, image-only foreign-language PDFs into fully formatted English documents (OCR + LLM translation + PDF reconstruction).
• Dark Phoenix — SaaS platform that turns full-length podcasts and videos into viral short-form clips for YouTube Shorts and TikTok, with active-speaker cropping and auto-captions.
• Edge — LunarTech's Learning Management System (LMS) powering courses, cohorts, and student progress.
• Imagine — AI-powered visual generation web app (imagine.lunartech.ai).
• Lens — LunarTech's media and tech-journalism department, publishing articles, editorials, podcasts, and video on AI, robotics, energy, and deep tech.
• Luna — LunarTech LABS product (in development).
• Memento — Hyperpersonalized AI learning companion (NotebookLM-style) that turns uploaded sources into chat, audio overviews, simulations, and custom multimedia.
• Octavia — Multilingual video and course translation/localization AI that dubs and localizes spoken content end-to-end across languages.
• Odyssey — Intelligent large-scale document home — a repository and knowledge platform for organizing and querying documents.
• Phoenix — A fleet of 100+ AI agents for content creation, research, course generation, SEO, and automated knowledge work.
• Simon — Self-hosted social-media publishing system (Postiz-based) with working PDF uploads and LinkedIn document/carousel posting.
• SilverAI (Silver) — Autonomous Knowledge Engine that researches and writes high-stakes professional documents, scaling from 1-page summaries to 1,000-page manuals with a deep-research agent.
• Superpowers — Complete agentic software-development workflow built from composable "skills" that give coding agents spec-driven, test-driven, subagent-orchestrated development.
• Trilogy — Platform connecting individual capital to private markets, opening private-equity and venture access historically reserved for institutional LPs.
• Valeria — AI-powered unified sales engine that replaces 10–15 disconnected tools, connecting enterprise knowledge, content, ads, outreach, inbound, and analytics in one system.

• Phoenix — Proprietary platform of specialized AI agents and LLMs for server-side content generation, research, course creation, and SEO — scaled to 300+ autonomous multi-modal agents that drove a 10M+ view increase in a single month with zero ad spend.
• Multimodal AI Agent — Integrated reasoning system that fuses video, audio, text, and image understanding into a single agent.
• Legal RAG System — Conversational AI for legal document analysis — a vector-based chat system over 10,000 British Columbia Real Estate Law documents, built with embedding models and Mistral AI.
• DeepRAG — End-to-end document intelligence system built on LangChain, Pinecone, and multi-model embeddings.
• Multi-Document Query Orchestration — LlamaIndex + OpenAI tool-routing engine that reasons across many documents at once.
• LunarTech Resume Evaluator — AI-powered candidate assessment and scoring system.
• YouTube Full-Stack AI Team — Automated, role-based content pipeline that researches, writes, and produces video content.
• Social Media Video Reposter — Multi-platform content automation for cross-posting and reformatting video.
• LinkedIn Content Generator — Viral post creation engine with 1.5M+ views generated.
• YouTube Shorts Generator — Automated short-form content creation pipeline.
• AI Web Journalist — Automated news repurposing and article generation engine.
• Webflow CMS Generator — Dynamic website and content automation system.
• Webflow CMS Updater — Scalable course-management and content-sync platform.
• LLM Fine-Tuning for YouTube — Contextual comment-intelligence models fine-tuned on channel data.
• Rachel — Enterprise-level voice assistant built in TypeScript.

• Hybrid AI Stroke Risk Pipeline — Clinical risk-prediction system combining structured data with LLM intelligence for stroke modeling.
• Damage Propagation Modeling for Aircraft Engines — ML pipeline (Random Forest, Linear Regression, LSTM, XGBoost) modeling engine wear and remaining useful life.
• Superstore Data Analysis — Predictive insights and business-intelligence visualization platform.
• NLP Movie Recommendation Engine — Production-grade recommender using count vectorization and cosine similarity.

• Advanced ChatGPT-Driven Algorithmic Trading System — Crypto, forex, and stock-market strategy engine in Python.

• Max (Genesis Edge AI) — Professional edge-AI quadcopter with carbon-fiber tube airframe, Jetson Orin compute, and modular mission payloads.
• Moose — 1600mm twin-motor 3D-printed fixed-wing UAV for long-range FPV and aerial mapping.
• Blackbee — High-speed 8S carbon-frame FPV quadcopter with servo-driven camera tilt.
• Whitebee — Sub-400g FPV speed quad engineered for 300+ km/h straight-line runs (313 km/h proven).
• Vixen — Compact 4-inch dead-cat pusher FPV drone with DJI O4 Pro integration.
• Wings — Flying-wing UAV platform in the Nomad lineup.
• Turbojet Engine Manufacturing — In-house turbojet engines in the 20–100 KGF thrust class.
• UAV Manufacturing — High-performance aerial systems for defense applications.
• 3D-Printed UAV Frames — Lightweight aerodynamic airframe structures.
• Carbon Fiber 3D Printing — High-strength frame fabrication for aerial platforms.

• Robotic Arm Manufacturing — Precision-engineering automation and motion systems.

Nomad designs and manufactures a full spectrum of unmanned aerial systems — from professional edge-AI mission platforms to record-chasing FPV speed machines. Every aircraft below ships with source CAD/STL files, assembly manuals, and frame-only variants.

A professional edge-AI quadcopter built around a T700 carbon-fiber tube airframe, glass-filled nylon enclosures, and onboard NVIDIA Jetson Orin compute (Nano up to 40 TOPS / NX up to 100 TOPS). It runs ARK Flow / PX4 avionics, carries up to 4 kg of modular dovetail-rail payload, flies up to 30 minutes on 15-inch carbon props, and targets defense R&D, autonomy research, mapping, and industrial inspection with NDAA-focused sourcing.

A 1600mm-wingspan, twin-tractor-motor, V-tail fixed-wing platform with removable wings, a swappable modular nose, and a central payload bay. 3D-printed in lightweight LW-PLA/LW-ASA and reinforced with carbon-tube spars, it's built for long-range FPV, aerial mapping, and terrain survey (2–4.5 kg AUW, 60–70 km/h cruise, Selig S3021 airfoil).

An advanced 8S high-speed FPV quad pairing a streamlined printed shell with a carbon main frame, 2312/1700kV power, a Fettec 8S 65A ESC, and a servo-driven camera tilt mechanism (up to ~90°). Built for aggressive forward-flight performance and serious DIY pilots.

A sub-400g FPV speed quadcopter on the Armattan Movr163 CP2 frame with aerodynamic printed fairings, a 6S T-Motor V2208/1950KV powertrain, and Walksnail HD FPV. Engineered for 300+ km/h speed runs — 313 km/h proven in spec testing.

A custom 3D-printed 4-inch FPV drone using a dead-cat + pusher geometry to clear the props from the camera view. Built around a 20×20 stack, DJI O4 Pro camera/VTX, ELRS, GPS, and a 6S powertrain, printed in PA-CF/PET-CF with TPU mounts and skids.

A flying-wing UAV platform in the Nomad lineup, available alongside its dedicated frame kit.

Each model is also offered as a bare frame: max-frame, moose-frame, blackbee-frame, whitebee-frame, vixen-frame, wings-frame.

Software engineer with a decade of professional experience architecting scalable systems and leading high-stakes technical initiatives. As the founder of LunarTech, I lead a multidisciplinary team focused on advancing the frontiers of Deep Tech, Healthcare, Aviation, and Propulsion Systems.

My work is characterized by a "Substance-First" philosophy — bridging the gap between sophisticated, human-centric design and robust, high-performance back-end architectures.

LunarTech operates at the intersection of academia and industry, engineering the tools and frameworks that define the next generation of technical excellence.

• LunarTech Academy — A global powerhouse reaching an audience of over 100 million, with 100+ deployed courses and multiple technical handbooks. An industry benchmark for high-level technical mastery.
• Flagship Bootcamps — Home to the world's most sought-after AI Engineering and Data Science bootcamps.
• LunarTech Edge — The Learning Management System (LMS) that delivers LunarTech's courses, cohorts, and student progress tracking.
• LunarTech Labs — The applied-AI lab behind Babel, Octavia, Phoenix, and the agent ecosystem.
• Babel & Octavia — Proprietary architectures that translate documents and video across languages while streamlining developer and learning workflows.
• Institutional Mission — Driven by human empowerment, we provide intensive academic support to remove barriers to professional entry for developers worldwide.

20+ technical handbooks and books authored across AI, engineering, and career development:

• The Lithography Handbook: Machines, Markets, and the Next Wave of Semiconductor Startups
• The Claude Code Handbook: A Professional Introduction to Building with AI-Assisted Development
• The Case for End-to-End Engineering Education: Preparing Institutions for a Dynamic Future
• Pioneering Next-Gen Healthcare with AI, Epigenetics, and Bioengineering
• AI in Agriculture: How AI-Enhanced Farming Can Increase Crop Yields [Full Book]
• How AI Agents Can Help Supercharge Language Models – A Handbook for Developers
• The Generative AI Handbook – How GenAI is Impacting Business and Innovation
• Next-Gen Large Language Models: The Retrieval-Augmented Generation (RAG) Handbook
• Applied Data Science with Python – Business Intelligence for Developers [Full Book]
• How to Build a Production-Grade Movie Recommender in Python – A Machine Learning Handbook
• How to Create a Great Personal Portfolio Page – a Handbook for Beginners
• Advanced Prompt Engineering for Content Creators – Full Handbook
• Python Back-End Development – Handbook for Beginners
• Personal Branding for Developers – A Step by Step Handbook
• Advanced Object-Oriented Programming in Java – Full Book
• How to Innovate as a Developer – a Roadmap to a Successful Tech Career [Full Book]
• The Java Interview Prep Handbook – 50 Questions Solved + Code Examples
• The Software Engineer Internship Handbook – How to Launch Your Coding Career
• Data Structures Handbook – The Key to Scalable Software
• Learn Java Fundamentals – Object-Oriented Programming [Full Book]

• Google Cybersecurity Professional Certificate — Google • Issued Dec 2024 • Credential ID: 89E8W6B6U3N5
  Covers Foundations of Cybersecurity, Assets/Threats/Vulnerabilities, Networks & Network Security, Linux & SQL, Python automation, Detection & Response, Security Risk Management, and Cybersecurity job readiness.
• Introduction to AI — Google • Issued Jul 2025 • Credential ID: LY4E3A7FEXI9
• Avionics and Aircraft Systems: Fundamentals — Coursera Instructor Network • Issued Jul 2025 • Credential ID: LDXCK147JK9V

"A committed Software Developer... featured in Forbes and Entrepreneur for his innovative and grounded approach to digital solutions."

Featured and covered in:

Forbes • Entrepreneur • Bloomberg • Insider • Benzinga

Partnered with and accepted into the startup / founder programs of the world's leading AI and technology companies:

• NVIDIA — Inception startup program (accelerated computing & edge AI).
• Microsoft — Startup / founders program (Azure, cloud, and AI tooling).
• Google — Startup / cloud program (GCP and AI infrastructure).
• Anthropic — Partner and builder program (Claude models and Agent SDK).
• OpenAI — Partner and startup program (frontier model access).

The systems above weren't built from a tutorial — each one forced a specific, hard-won skill. This is what they add up to.

• Agentic systems — designing multi-agent architectures with role decomposition, tool routing, planning/critique loops, and long-horizon autonomous execution.
• LLM application development — prompt engineering and prompt architecture, structured output, function/tool calling, system-prompt design, guardrails and multi-model "smart barrier" safety routing across Claude, GPT, and Gemini.
• Retrieval-Augmented Generation (RAG) — end-to-end pipelines with LangChain and LlamaIndex, vector databases (Pinecone), multi-model embeddings, chunking strategies, hybrid search, re-ranking, and multi-document query orchestration with tool routing.
• Fine-tuning & model adaptation — supervised fine-tuning of LLMs on domain data (YouTube comment-intelligence models), dataset curation, and evaluation.
• Agentic OCR & document AI — combining OCR with LLMs to translate and reconstruct scanned, image-only PDFs while preserving tables, figures, and layout (Babel, Babel Extreme, Clarke).
• Multimodal AI — fusing video, audio, text, and image understanding into single reasoning systems; speech-to-text (WhisperX), active-speaker detection, and AI video/voice generation and dubbing (Octavia, Dark Phoenix).

• Classical & deep ML — Random Forest, Linear/Logistic Regression, XGBoost, LSTM, and neural networks for prediction and time-series (aircraft engine damage propagation, stroke-risk modeling).
• Frameworks — TensorFlow, PyTorch, scikit-learn, ONNX, TensorRT, OpenCV.
• NLP — count vectorization, TF-IDF, cosine similarity, semantic search, recommendation systems.
• Data engineering & analytics — preprocessing, feature engineering, scaling, predictive modeling, and BI visualization (Superstore analysis); hybrid pipelines that blend structured data with LLM intelligence for clinical risk.

• Languages — Python, TypeScript, JavaScript, Java, C/C++, Kotlin, SQL.
• Backend — scalable, distributed back-ends; REST/gRPC APIs; queue systems (Inngest); auth, billing, and credit/subscription systems (Stripe); Convex, Supabase, Postgres.
• Frontend — Next.js, React, modern component systems and design-driven UI/UX.
• Voice & real-time — enterprise voice assistants (Rachel) and voice-AI agent integration.

• Content automation at scale — automated pipelines for YouTube, Shorts, viral clip generation, social reposting, and a LinkedIn engine that drove 1.5M+ views.
• CMS & web automation — programmatic Webflow CMS generation and updating for dynamic sites and course platforms.
• Self-hosting & DevOps — Docker, Linux, deployment/debugging of self-hosted stacks (Postiz/Simon), MIME/upload/API-level patching, and platform integrations (LinkedIn API).
• SaaS productization — turning models into shipped products with pricing, queues, payments, and dashboards.

• Cloud platforms — AWS, GCP, Azure (incl. their startup/AI programs), Databricks.
• Edge AI — NVIDIA Jetson Orin (Nano/NX), on-device inference, TensorRT optimization, real-time perception on constrained compute.
• MLOps & DevOps — model serving, GPU-accelerated processing (FFMPEGCV), pipeline orchestration, CI/CD, system/infrastructure monitoring, alerting, and reliability at load.

• Defensive security — network security, threat detection, vulnerability management, and security risk management (Google Cybersecurity Professional certified).
• DevSecOps — security automation with Python, Linux/SQL for security, detection-and-response workflows, and secure self-hosted deployments.
• Applied to products — multi-model "smart barrier" safety routing and zero-tolerance content-integrity architecture across foundation models.

• Performance marketing — paid-ad strategy across Google, Facebook, Reddit, and Instagram Ads (9% Reddit CTR), with campaigns driving millions of mobile-game installs (Clash of Clans, Clash Royale).
• Organic growth — zero-cost growth engines reaching 10,000 new users/month and 10M+ views in a month; social strategy lifting traffic 30% and engagement 50%.
• Design & media — UI/UX, graphic design, and video/thumbnail production with Figma, Photoshop, and the Adobe Creative Suite.

• UAV design & manufacturing — full aircraft design across multirotor, fixed-wing, and flying-wing platforms; airframe architecture, powertrain selection, and flight-controller integration (PX4, ArduPilot, Betaflight, MAVLink/MAVROS).
• Propulsion — turbojet engine manufacturing in the 20–100 KGF thrust class; propulsion-system architecture.
• Avionics & autonomy — GNSS/RTK, telemetry/RF links (ELRS, Crossfire, RFD900), companion-computer autonomy, waypoint/survey missions, and onboard AI payloads.
• Advanced manufacturing — CAD/STEP design, 3D printing (LW-PLA/PETG/TPU/PA-CF), carbon-fiber fabrication, and precision robotic-arm automation.

• Technical leadership — founding and running multidisciplinary teams across AI, aviation, defense, and education; architecting systems others build on.
• Education & communication — 20+ published handbooks/books and 100+ courses; the rare ability to take something deeply technical and make it teachable.

• Direct: vahe@lunartech.ai
• Corporate: lunartech.ai
• Portfolio: vaheaslanyan.com