Curriculum Vitae

calendarDec 13, 2025 · 7 min read

Professional Summary

Senior engineer specializing in production multi-agent LLM systems and distributed blockchain protocols. 3+ years architecting custom agentic frameworks with RAG, tool orchestration, and memory systems. 4+ years building Cosmos-based protocols with hands-on validator operations. Expert in Python and Go with deep experience building autonomous agents and distributed systems from POC to production.

Recent Work:

  • Built multi-agent LLM POC from personal project (2023), then contributed to commercial production system serving thousands of users (2025)
  • Core engineer and top-10 validator operator for Cosmos-based blockchain protocol
  • Developed specialized agents, implemented RAG with real-time blockchain data, built benchmarking infrastructure

Background: 25+ years across biometric systems R&D (face/fingerprint/iris recognition algorithms), autonomous robotics (designed and implemented bio-inspired neural network frameworks from scratch in Assembly/C/C++/C#), electronics design (128-channel parametric ultrasound generators, multi-head 3D printer controllers), network automation at scale, real-time human tracking systems, embedded systems.

Technical Depth: Agentic AI from scratch in Go, hand-crafted RAG solutions, token usage optimization, distributed consensus, Byzantine fault tolerance, protocol design, low-level neural network implementation, computer vision, sensor fusion, real-time systems, hardware/software integration, signal processing, performance optimization at every level.

Differentiator: Rapid system absorption + cross-domain pattern recognition from bare metal to distributed protocols. I understand systems at every abstraction level and identify non-obvious issues by applying patterns across domains.

Experience

Synternet — Senior Engineer

OCT 2020 - DEC 2025

Agentic AI Engineering (2023-2025)

2025 Production System (Design Contributor & Agent Developer):

  • Contributed to commercial multi-agent LLM bot serving thousands of users with Telegram and web interfaces - company's flagship product
  • Built comprehensive LLM benchmarking infrastructure and tooling for model evaluation across multiple providers
  • Developed specialized agents and contributed to custom agentic framework design in Python (built from scratch after evaluating LangChain/LlamaIndex)
  • Primary responsibility: real-time blockchain data integration and Telegram game integration with bot system
  • Implemented agents utilizing RAG pipeline with PostgreSQL vector database, tool-calling, and real-time blockchain data feeds
  • Contributed to prompt engineering for multi-step reasoning and event relevance scoring using 1M token context models
  • Contributed to comprehensive instrumentation analyzing p95 latencies, identifying failure causes and optimizing reasoning times
  • Participated in framework evolution toward graph execution (LangGraph-style) with event-driven agent orchestration

2023 POC (Sole Architect & Developer):

  • Architected and built multi-agent system in Go from scratch, adapting personal mental health chatbot framework to company's blockchain data layer
  • Designed modular architecture with agents as distributed services communicating via NATS message bus using Protobuf/JSON serialization
  • Implemented vector database from scratch in SQLite with hand-coded similarity metrics (cosine, euclidean) to understand retrieval mechanics
  • Built memory systems: per-user short-term (context management with summarization) and user/global long-term (vector similarity search for semantic retrieval)
  • Integrated OpenAI API with agentic tool usage patterns and multi-agent coordination
  • Deployed for internal team exploration of semantic search and agentic capabilities - demonstrated viability of multi-agent approach for blockchain data

Technical Challenges: Multi-agent orchestration, RAG optimization for real-time blockchain data, prompt engineering for reasoning tasks, memory system design, LLM provider reliability, event-driven agent communication, vector similarity implementation

Protocol Engineering & Blockchain Infrastructure (2020-Present)

Protocol Development:

  • Engineered Cosmos-based blockchain protocol from ground up (Go, Tendermint consensus, ABCI integration, state machines)
  • Designed and implemented decentralized DARP routing protocol for distributed topologies (patented)
  • Engineered high-throughput cross-chain bridges with cryptographic security guarantees
  • Developed Byzantine fault tolerance implementations and consensus algorithm optimizations
  • Integrated network layer with application layer for transaction ordering and finality
  • Built high-throughput data pipelines and real-time streaming systems using NATS

Infrastructure & Validator Operations:

  • Operating top-10 validator with 99.9% uptime managing $70M+ staked value
  • Managed network upgrades, governance proposals, and consensus parameter changes
  • Built monitoring and incident response systems for 24/7 validator operations
  • Led DevOps training program for distributed validator network
  • Engineered secure Ethereum→Cosmos bridge with zero security incidents

API Integration & Data Layer:

  • Developed Go SDK library to interface with the data layer, JWT management, Protobuf, JSON
  • Integrated 15+ APIs including blockchain protocols (Ethereum, Solana, Cosmos, Infura, Etherscan), market data (CoinMarketCap, CoinGecko), and communication platforms (Telegram, Slack)
  • Built Telegram mini-game bot integrating Web3 functionality with real-time blockchain data

Technical Challenges: Byzantine fault tolerance implementation, cross-chain asset bridges, distributed state synchronization, consensus algorithm optimization, real-time data streaming at scale

Stack: Python, Go, OpenAI API, OpenRouter, RAG, vector databases (SQLite, PostgreSQL pgvector), NATS, Protobuf, Telegram Bot API, Slack API, multi-agent systems, prompt engineering, token optimization, LLM orchestration, Cosmos SDK, Tendermint, IBC, ABCI, gRPC, Ethereum, Solidity, Docker, Kubernetes, Git, GitLab, JIRA, Confluence, Miro, Ansible

Genus AI — Data Engineer

AUGUST 2019 - SEPTEMBER 2020

API Integration & Backend Development:

  • Integrated ML inference APIs into user-facing React application with real-time prediction pipelines
  • Built Facebook Graph API integration for advertising platform with OAuth flow implementation, webhook handling, and rate limit management
  • Designed GraphQL API layer (Django/Python) connecting ML models to frontend with optimized query resolution
  • Implemented comprehensive error handling and retry logic for third-party API calls with exponential backoff
  • Achieved 100% test coverage using TDD - extensive unit and integration testing for API interactions
  • Maintained production monitoring via CloudWatch with alerting for API failures and latency issues

Technical Challenges: ML model serving at scale, third-party API reliability, async job processing, GraphQL query optimization, serverless cold starts

Stack: Python, Django, GraphQL, REST APIs, Facebook Graph API, AWS (Lambda, Fargate, S3, SageMaker, CloudWatch), React, OAuth 2.0, TDD, Git, Phabricator, Arcanist

Barclays — Automation Engineer & Scrum Master

SEPTEMBER 2016 - APRIL 2019

Integration & Automation:

  • Designed network automation solutions integrating enterprise infrastructure APIs with Python orchestration layer
  • Built REST API integrations with ServiceNow, MySQL, and network device APIs for automated provisioning
  • Implemented async task processing with Celery and RabbitMQ for long-running automation workflows
  • Developed robust error handling and rollback mechanisms for critical infrastructure changes
  • Led code reviews focusing on API integration best practices, error handling, and OOP design patterns
  • Led Scrum ceremonies and facilitated team collaboration

Stack: Python, Django, Flask, Celery, RabbitMQ, MySQL, REST APIs, ServiceNow API, network automation, Git, JIRA, Confluence

CTO | Phantom Corp | Jun 2015 - Dec 2016

  • Developed revolutionary multi-head 3D printer technology
  • Applied color recognition algorithms enabling camera-integrated color 3D printing
  • Designed 3D printing algorithms optimizing multi-material workflows
  • Combined robotics, computer vision, and embedded systems expertise

Technical Stack: C/C++, computer vision, robotics control systems, embedded firmware

Neurotechnology — AI/Biometrics R&D Analyst

MARCH 2004 - AUGUST 2016

AI/ML & API Development:

  • Developed biometric recognition SDKs and APIs (fingerprint, face, iris) serving enterprise customers
  • Conducted AI research on neural network algorithms for real-time pattern recognition
  • Built bio-inspired neural networks from scratch for autonomous robot navigation (C/C++/Assembly)
  • Designed RESTful APIs and SDK interfaces for biometric system integration
  • Implemented real-time video processing pipelines with RTSP protocol integration
  • Remote iris scanning technology requiring precise hardware/software API integration

Stack: C/C++, C#, Python, computer vision, machine learning, neural networks, OpenCV, API design, RTSP, Git, SVN

Freelance Engineer | Self-Employed | Dec 1998 - Present

Selected Projects:

  • Multi-agent Mental Health Chatbot (2022-2024): Personal project using OpenAI API - architectural foundation adapted for Synternet's 2023 blockchain data POC
  • Art Ticket printing: CUSOM thermal printer integration into a larger ticket system
  • Autonomous Robot Navigation: Custom AI system for robot navigation (KTU competition participant) - C/C++, Python, Go
  • Remote Controlled Reflow Station: - Custom auto calibrating SMD Reflow controller FW and SW. ESP32.
  • Weather Station: - Streaming air quality parameters over MQTT: diagnosing Office Spaces, and a Museum air quality over extended periods of time
  • Spectral Analysis Systems: FFT signal processing for Vilnius University physics research
  • Specrometer control and signal acquisition: For Liquid Crystal research lab
  • Color Science R&D: Custom colorimeter design and calibration (multi-LED/multispectral), device integration and visualization (Python, Go), explored color space conversions, color matching algorithms, and colorimetric data analysis for experiments and art installations
  • Wax Printing Art: Printing head controller for Wax art installation
  • 2D Relaxation Time Calculation Algorithm: For physics research (Masters/PhD work)
  • Cryostat Control Software: With ultra-low frequency spectral analysis capabilities
  • Face Recognition Art Installation: "Speaking Doors" project with real-time recognition
  • 128-Channel Parametric Ultrasound Generator: Electronics design for parametric audio and ultrasound levitation
  • Camera LUT calibration: Custom Hardware Camera module LUT calibration firmware
  • Chronograph: For ballistic velocity measurements
  • POV Display: And 16-channel stepper controller
  • Disco Light System: WinAMP plugin, FFT processing, LPT interface

Education

  • Ph.D. Candidate | Vilnius University | 2008-2014
  • M.Sc. Telecommunication Physics & Electronics | Vilnius University | 2006-2008
  • B.Sc. Business Management | Vilnius University | 2007-2008
  • B.Sc. Telecommunication Physics & Electronics | Vilnius University | 2002-2006