ENGINE_OS::NEXUS_v1.4_ACTIVE
NYC / NJ12:00:00 EST

Architecting Autonomous Intelligence & Production AI Swarms.

I am William Volodarsky — an Agentic AI Engineer & Systems Orchestrator (Fordham CS, Magna Cum Laude). I build local-first multi-agent systems using LangGraph, deterministic context injection, and human-in-the-loop governance to replace fragile AI demos with reliable production infrastructure.

Autonomous Swarm
4 Agents

LangGraph Router, Librarian, Career & Email subgraphs

Deterministic Injection
99.8%

Zero-cost DPFH directory injection vs chunk fragmentation

Deployed Automation
12 Tools

HL7 CDA XML, ADB Scraper, Playwright PDF, IMAP OAuth

Foundation Model Evals
300+ Hrs

Multi-axial alignment, LLM-as-a-Judge test harnesses

FLAGSHIP ARCHITECTURE

Nexus: The Multi-Agent Cognitive OS

Replacing single-prompt toy demos with a stateful, local-first multi-agent swarm. Nexus orchestrates health, career, and research workflows across 500+ structured markdown notes with zero hallucinated file structures.

DPFH vs. Traditional Vector RAGZero LLM Cost

Side-by-side operational architecture for structured knowledge retrieval.

Traditional Chunked Vector RAG
  • Chunk Fragmentation: Rips markdown tables, frontmatter, and headings into arbitrary 500-token slices.
  • Lost in the Middle: Semantic search retrieves out-of-context chunks without hierarchical awareness.
  • Hallucinated Paths: Generates plausible but nonexistent markdown note links.
Deterministic Pre-Flight Hydration
  • Zero-Cost Injection: Deterministic Python pre-node injects live disk directory tree into prompt.
  • Whole-Document Reading: Librarian traverses exact notes and structured wiki-links natively.
  • 100% Grounded Links: Every citation is a verified, live Obsidian deep link.
nexus/orchestration/dpfh_hydrator.pyDeterministic Hook
def hydrate_agent_context(state: SwarmState) -> dict:
    # 0 LLM tokens: Pure Python tree mapping & frontmatter injection
    live_tree = vault_reader.get_domain_structure(domain=state["domain"])
    curated_prompt = build_system_prompt(
        subconscious_rules=RULES_CONFIG,
        live_taxonomy=live_tree,
        conversation_summary=state["summary"]
    )
    return {"hydrated_prompt": curated_prompt}

Production Domain Tool Suite

12 purpose-built automation pipelines powering Nexus agents.

View Source on GitHub
HL7 CDA XML ParserClinical Health

Longitudinal clinical record synthesis & lab trend tracking.

11-Platform ADB ScraperMobile Ops

Android screen capture with multi-scroll OCR deduplication.

Dual Resume EngineDocument Systems

Headless Playwright PDF & python-docx with page-fill metric audits.

Google OAuth2 IMAP SyncCommunication

Asynchronous email ingestion & priority triage pipeline.

LLM-as-a-Judge HarnessEvaluation

12-case golden dataset test runner with JSON scoring reports.

Monaco HITL Diff QueueGovernance

Pre-commit side-by-side file write review in SQLite.

Trafilatura Web DistillerContent

Deterministic article cleaning & atomic Zettelkasten synthesis.

TTS Audio Podcast GeneratorAudio

Automated daily briefing audio synthesis via voice models.

Whisper Voice TranscriberMultimodal

Local asynchronous audio capture & thought triage.

Telegram Bot Swarm RelayEdge Capture

Mobile prompt execution with user-ID crypt-security.

Tree-Based Vault NavigatorFilesystem

Zero-cost directory tree & frontmatter metadata traversal.

Git-Crypt Secret EnclaveSecurity

Zero-knowledge client-side encryption for personal records.

INTERACTIVE SYSTEM PIPELINE

End-to-End Execution Architecture

Click any node in the execution DAG below to inspect how Nexus ingests, classifies, hydrates, evaluates, and safely commits autonomous agent actions.

Intent Classification & Conditional Dispatch

LangGraph Content Router

Deterministic zero-shot classifier matches user intent against domain agents (Health, Career, Email, Librarian). High-confidence matches route directly; complex cross-domain queries dispatch to the Librarian subgraph.

Core Technologies & Protocols:
LangGraph StateGraphConditional EdgesStructured Pydantic
EXECUTION_FRAMEWORKGROUNDED
def route_query(state: SwarmState) -> str:
    domain = classify_intent(state["messages"][-1].content)
    # Returns "career", "email", "librarian", or "general"
    return f"enter_{domain}_subgraph"
TRACK RECORD & LEADERSHIP

Professional Experience & Execution

Combining rigorous foundation model evaluation with 8 years of operational leadership, revenue scaling, and mission-critical compliance.

AI Training Fellow

Foundation Model Alignment
Handshake AI FellowshipRemote / New York, NY
Nov 2025 – May 2026
  • Performed iterative resampling with corrective feedback across LLMs and latent diffusion models.
  • Evaluated model outputs across multi-axial quality rubrics: factual accuracy, instruction following, safety, and non-deterministic code correctness.
  • Identified systematic failure modes (hallucination patterns, lost-in-the-middle context degradation), translating edge cases into actionable fine-tuning constraints.
LLM AlignmentRLHF EvaluationPrompt EngineeringHallucination AuditingError Taxonomy

Co-Owner & Operations Lead

2x Revenue Growth
SiliSlick KnivesFarmingdale, NY / Global
June 2017 – June 2025
  • Identified high-margin product opportunities through sales velocity and competitive data analysis, deploying new product lines that doubled private brand revenue over 18 months.
  • Engineered multi-channel international distribution (Amazon, Wholesale, D2C) across US, UK, and EU markets.
  • Managed regulatory compliance, international customs logistics, product design, marketing campaigns, and multi-tier supplier negotiations.
Operations LeadershipSupply ChainData AnalysisRevenue OptimizationRegulatory Compliance

Medical Records & Patient Services

100% HIPAA Compliance
Vitreous Retina Macula ConsultantsNew York, NY
June 2015 – Feb 2017
  • Co-managed clinical records for a premier high-volume ophthalmology practice, maintaining 100% HIPAA compliance across document handling, retrieval, and clinical transfers.
  • Built interactive tracking spreadsheets and scheduling databases for patient service quality, driving measurable improvements in clinical throughput.
HIPAA ComplianceClinical Data SystemsProcess OptimizationEHR Systems

Education & Academic Honors

Sept 2023 – Dec 2025Magna Cum Laude (GPA: 3.7)

Bachelor of Science in Computer Science

Fordham UniversityNew York, NY

Dean's List. Focus on Algorithms, Operating Systems, Database Systems, Artificial Intelligence, and Software Engineering.

Jan 2020 – Dec 2022Dean's List, Phi Theta Kappa (GPA: 3.7)

Associate of Science in Computer Science

LaGuardia Community CollegeLong Island City, NY

Rigorous foundation in Object-Oriented Programming (Java/C++), Data Structures, and Discrete Mathematics.

TECHNICAL PROFICIENCIES

Skills & Systems Competencies

Strictly production-tested competencies across agent orchestration, backend services, evaluation testbenches, and data pipelines.

LangGraph
agenticProduction
Multi-Agent Swarms
agenticProduction
Deterministic Pre-Flight Hydration
agenticSpecialty
ReAct Tool-Calling
agenticProduction
LangChain & LangSmith
agenticAdvanced
Structured Pydantic Outputs
agenticProduction
Dynamic Prompt Injection
agenticAdvanced
SSE Trace Streaming
agenticProduction
Vector RAG (ChromaDB, HyDE)
agenticAdvanced
OpenAI API & Anthropic
agenticAdvanced
Whisper Voice Processing
agenticProduction
Python
fullstackExpert
TypeScript
fullstackAdvanced
Next.js 16 (App Router)
fullstackAdvanced
React 19
fullstackAdvanced
FastAPI
fullstackAdvanced
Tailwind CSS v4
fullstackAdvanced
SQL (Postgres, SQLite)
fullstackAdvanced
Java & C++
fullstackProficient
Node.js
fullstackProficient
Pytest Suite
fullstackAdvanced
Pydantic V2
fullstackAdvanced
PowerShell & Bash
fullstackAdvanced
Playwright Headless
infraProduction
ADB / Android Automation
infraProduction
HL7 CDA Medical XML
infraSpecialty
Git / GitHub Actions
infraAdvanced
Git-Crypt Security
infraProduction
Telegram Bot API
infraProduction
Google OAuth2 & IMAP
infraProduction
Syncthing Mesh Sync
infraAdvanced
Linux / POSIX
infraAdvanced
Obsidian / Zettelkasten
infraMastery
LLM-as-a-Judge Eval
evalsSpecialty
HITL Two-Phase Commit
evalsSpecialty
Hallucination Auditing
evalsExpert
Context Degradation QA
evalsExpert
HIPAA Clinical Compliance
evalsExpert
LIVING DOCUMENT

Master Resume & Evidence Locker

Synchronized directly with production engineering and operations track record.

Request PDF / DOCX

William Volodarsky

Agentic AI Engineer & Systems Orchestrator

Fort Lee, NJ 07024 • bill@williamvolodarsky.com • github.com/brolodarsky • linkedin.com/in/williamvolodarsky

BS CS Magna Cum Laude (3.7 GPA)8 Yrs Ops Leadership

01 // Professional Summary

Agentic AI Engineer (CS, Magna Cum Laude) and creator of Nexus, a production-grade Agentic OS featuring a multi-agent LangGraph pipeline with Deterministic Pre-flight Hydration, a full-stack Next.js + FastAPI control panel with HITL governance, an LLM-as-a-Judgeevaluation framework, and 12 deployed automation tools. Eight years of business operations leadership provides the systems judgment to ship robust, compliant AI systems that solve real-world problems.

02 // Flagship Project: Nexus (Agentic OS)

Multi-Agent Orchestration Pipeline: Built a LangGraph content router that classifies incoming queries and dispatches to domain-specialized agents via conditional edges. The Career Agent implements Deterministic Pre-flight Hydration — a zero-cost Python orchestration node injects live vault file listings and declared dependencies into the system prompt before each LLM call, eliminating hallucinated context.
Filesystem-Native Retrieval (Librarian Agent): Replaced an earlier ChromaDB Vector RAG pipeline with a LangGraph ReAct agent traversing 200+ notes via four filesystem tools (read_toc, read_note, search_vault, get_vault_structure) with tree-based directory injection and frontmatter filtering.
Full-Stack HITL Control Panel: Engineered a Next.js + Tailwind CSS dashboard backed by FastAPI for real-time agent fleet monitoring, conversational search, and a Monaco Editor side-by-side diff viewer for pre-commit review of autonomous agent writes — backed by a SQLite transaction queue.
Evaluation & Observability: Built an automated LLM-as-a-Judge eval framework grading responses on accuracy, groundedness, and hallucination detection via a 12-case golden dataset with structured JSON reports and append-only JSONL run logs.

03 // Professional Experience & Leadership

Handshake AI Fellowship — AI Training Fellow
Nov 2025 – May 2026

Performed iterative resampling with corrective feedback across LLMs and latent diffusion models; evaluated outputs across multi-axial quality rubrics, identifying systematic failure modes (hallucination patterns, context degradation) and translating edge cases into actionable fine-tuning constraints.

SiliSlick Knives — Co-Owner & Operations Lead
June 2017 – June 2025

Doubled private brand revenue over 18 months through sales velocity analysis and high-margin product expansion. Managed multi-channel international distribution (Amazon, Wholesale, D2C) across US/UK/EU markets, overseeing regulatory compliance, customs logistics, and supplier negotiations.

04 // Education & Honors

Fordham UniversityBS in Computer Science, Magna Cum Laude (GPA: 3.7)
Dean's List • Sept 2023 – Dec 2025
SYSTEM_STATUS // OPEN FOR OPPORTUNITIES

Let's Build Production AI Together.

I am currently evaluating high-impact Agentic AI Engineering and Systems Orchestration roles (NYC, Hybrid, or Remote). Whether you are building autonomous agents, tackling LLM alignment, or scaling complex operations, let's connect.

Official Communication Channel
bill@williamvolodarsky.com
Send Email