Vassiliy Lakhonin

I build evidence-discipline tooling for strategic-risk AI agents: Agenda Intelligence, an open-source runtime (one core over MCP, HTTP, A2A, and Cloudflare Workers) and a four-repo skill portfolio for structured sanctions, corridor, geopolitical, energy, trade, and regulatory analysis — with six live A2A workers and three interactive demos (Deal Risk Gate, Market Entry Readiness, CIS exposure). Grounded in donor compliance, audit readiness, and cross-border program delivery. Separately, GrantFlow for agent-native grant workflows.

Almaty, Kazakhstan (UTC+5) · Available for Central Asia, MENA, Europe, and U.S. overlap roles. · Last updated: 2026-06-05.

Run Deal Risk Gate Run Market Entry Gate Run CIS Exposure Gate For analysts For recruiters LinkedIn

Middle Corridor deal risk

Evidence-readiness triage before a Kazakhstan / Caspian transit deal moves toward signature.

Start Deal Risk Gate

Kazakhstan market entry

Distribution, import, service, showroom, EPC, renewable-energy, infrastructure, and partner-entry readiness.

Start Market Entry Gate

CIS counterparty exposure

Counterparty EDD-readiness triage for CIS / Caucasus / Central Asia secondary-sanctions exposure.

Start CIS exposure check

Analyst / policy-risk review

Agenda Intelligence repos, evidence discipline, live workers, and case-study trail.

Open analyst route

Hiring or recruiter screen

Role fit, verified outcomes, ATS-friendly resume, and machine-readable evidence surfaces.

Open recruiter route

Agent-readable profile

Canonical rules, JSON profile surfaces, and LLM/ATS evaluation instructions.

Open agent discovery

Agenda Intelligence — open-source strategic-risk skill portfolio

A four-repo system for AI agents producing auditable strategic-risk analysis — not generic geopolitical summaries. Built around evidence discipline (per-claim Axis A/B provenance, four labelled evidence modes, explicit unknowns) and mechanism-first reasoning. Aimed at fund analysts, compliance and risk teams, journalists, policy and trade advisors, and AI builders embedding regional intelligence into agents.

Horizontal reasoning skill — Global Think Tank Analyst: how an agent should structure a policy-risk memo across sanctions, regulatory, geopolitical, and trade questions.
Vertical specialist: Central Asia + Caspian — central-asia-caspian-hybrid-intelligence-skill: sanctions, AML, banking, corridor and energy risk across Kazakhstan, Uzbekistan, Azerbaijan and the Caspian.
Vertical specialist: Gulf + Middle East — gulf-middle-east-hybrid-intelligence-skill: Iran sanctions, GCC banking, sovereign wealth, energy markets, Hormuz / Bab-el-Mandeb maritime risk.
Product entry point + evidence-discipline layer — Agenda Intelligence MD: MCP server with the analyze tool (geography routing, structured request and memo schemas, optional Anthropic API path), plus the validation surface — schemas, evidence audit, scoring, CLI / MCP / CI tooling.

The reasoning skills define how agents reason. Agenda Intelligence MD defines how the output is audited. Together they let agents produce auditable strategic intelligence, not plausible-sounding summaries.

Live A2A agents

Public, no-payment A2A/JSON-RPC endpoints that put the open-source skill stack to work as live triage and routing surfaces. Full analysis still runs through the installable Agenda Intelligence MD MCP package. Boundary, stated on each agent card: no legal, financial, compliance, investment, or trading advice; no autonomous live source retrieval; no factual-truth verification.

Agenda Intelligence A2A — generic strategic-risk triage and source planning across the portfolio's frames. Endpoint: JSON-RPC. Listing: Agenstry.
Kazakhstan / Middle Corridor Deal Risk Gate — region-specific deal-risk triage built on the Central Asia + Caspian frame: sanctions exposure, Middle Corridor and Trans-Caspian logistics disruption, policy signals, source gaps, watch-next indicators. Try the interactive demo. Endpoint: JSON-RPC. Listing: Agenstry.
Kazakhstan Market Entry Readiness Gate — distribution, import, service, showroom, EPC, renewable-energy, infrastructure, technology-transfer, and partner-entry evidence triage: gate decision, readiness label, evidence gaps, claim audit, owner actions, and watch-next indicators. Try the interactive demo. Endpoint: JSON-RPC.
CIS Secondary-Sanctions Exposure — counterparty exposure evidence triage for CIS-domiciled counterparties (OFAC / EU / UK list context, ownership and transit gaps, decision-readiness). Try the interactive demo. Endpoint: JSON-RPC.
Agentic Interaction Trust Gate — evidence triage for agent-mediated actions (identity, authorization, tool scope, intent) before a high-stakes action runs. Endpoint: JSON-RPC.
Gulf Maritime Exposure Gate — evidence triage of maritime sanctions and chokepoint-disruption exposure for a vessel or voyage through the Strait of Hormuz, Gulf of Oman, Bab-el-Mandeb, or Red Sea (Iran-oil, Russia price-cap, dark-fleet, STS transfer, P&I gap; exposure signal, evidence gaps, chokepoint disruption watch). Built on the Gulf + Middle East frame. No vessel-ownership resolution or identity verification. Endpoint: JSON-RPC.

All five are JWS-signed and live on Cloudflare Workers; this site's ownership is verified on Agenstry. No paying customers — portfolio-grade, illustrative usage only.

Contact: vassiliy.lakhonin@gmail.com · linkedin.com/in/vassiliy-lakhonin
Last updated: 2026-06-05 · freshness.json for per-artifact timestamps.

Summary

Program/Portfolio/PMO leader focused on evidence, compliance, and cross-border delivery across development and private-sector settings. I manage multi-country portfolios, risk, stakeholder alignment, budget monitoring, and vendor oversight with an evidence-first approach. Languages: Russian (native), English (professional working proficiency).

Portfolio status

All four repos are public and actively maintained. “Validation” is not one thing here — each repo states which honest maturity framework applies.

Vertical specialists — agent-validated against their own canon. Central Asia + Caspian Hybrid Intelligence Skill and Gulf + Middle East Hybrid Intelligence Skill clear their two-bar Definition of Done: source-anchored example majority plus self-scored agent-eval delta cases, evidence-mode mapping through Agenda Intelligence analyze, and honesty discipline. All structural deltas are self-scored — not factual verification, not model-quality comparison, not aggregate benchmark. Practitioner review is an optional, audience-gated layer, not yet done. See each repo’s STATUS.md.
Horizontal reasoning skill — practitioner-review-first framework. Global Think Tank Analyst uses its own VALIDATION_PLAN.md, because horizontal reasoning across many domains does not map onto the vertical source-anchored-majority criterion. One self-scored agent-eval delta case committed; zero external practitioner reviews recorded.
Product runtime — v1.0.x contract frozen. Agenda Intelligence MD is the schemas / validation / evidence-audit / scoring / CLI / MCP / HTTP / A2A layer the other three repos compose with. The v1.0 request/memo contract is frozen (ADR 0003); changes ship additively under that freeze. It hosts five vertical workers (Middle Corridor Deal Risk Gate, Kazakhstan Market Entry Readiness Gate, CIS Secondary-Sanctions Exposure, Agentic Interaction Trust Gate, Gulf Maritime Exposure Gate), deployed on Cloudflare Workers and JWS-signed.

No production-usage, adoption, paying-customer, or benchmark numbers are claimed across any of these repos.

Work experience

Regional Monitoring and Evaluation (Performance) Manager — DAI (Almaty, Kazakhstan)

Apr 2023 – Apr 2025 Managed performance monitoring, reporting, data quality, and verification systems for a USAID-funded portfolio across four Central Asian countries.

Key results:

Delivered 100% on-time donor reporting for 8 consecutive quarters.
Reduced partner data submission delays by 40% through structured reporting and QA processes.
Strengthened M&E and reporting capacity across 15+ local partners through templates, coaching, and data checks.
Maintained audit-ready evidence, verification, and KPI documentation.
Supported two USAID audits with zero findings through evidence preparation and data validation.
Coordinated KPI tracking across four country offices and HQ to support performance monitoring.
Introduced AI-assisted reporting workflows, reducing report preparation time by 25%.

Independent Consultant — Program Design and Delivery Operations

Apr 2021 – Mar 2023 Delivered fixed-scope consulting assignments for nonprofit and private-sector clients, supporting the full project cycle from concept development and partner coordination to capacity-building design, implementation oversight, documentation control, and operational follow-up.

Project Management Specialist / Training Consultant — DAI (Almaty, Kazakhstan)

Mar 2019 – Feb 2021 Supported delivery of a USAID-funded business enabling environment portfolio across five Central Asian countries, contributing to stakeholder engagement, risk tracking, donor compliance, and implementation readiness.

Key results:

Delivered 25+ project outputs on time across scopes of work, assessments, trainings, compliance reports, and operational documentation.
Managed timelines, budget tracking, partner follow-up, and stakeholder coordination across the region.
Coordinated with 5 implementing partners and 50+ private-sector organizations.
Supported donor technical audits in 2019 and 2020 with complete evidence packs, including budget controls, deliverable sign-offs, risk logs, and partner documentation.
Used structured risk registers and escalation procedures to manage schedule, dependency, and delivery risks.
Co-developed and delivered Training-of-Trainers streams for a USAID-funded regional program, enabling cascade delivery to 100+ SME beneficiaries.
Supported training design, partner coordination, participant follow-up, workshop logistics, documentation, and reporting across Central Asia.

Researcher & Research Expertise Services — Multiple Clients

Jan 2014 – Dec 2018 Delivered research and project support across policy-focused assignments for academic institutions, think tanks, and nonprofit organizations.

Conducted desk research, literature reviews, fieldwork support, translation, and database setup for multi-country studies.
Contributed to policy briefs and peer-reviewed outputs by organizing evidence and synthesizing research inputs.
Supported workshops, conferences, participant coordination, logistics, and facilitation.

Education

Master in Politics and Security (Central Asia), OSCE Academy in Bishkek — 2013
Bachelor in Sociology (Dual-degree with Bard College, NY, U.S.), American University of Central Asia — 2012

Professional development

Future of AI — BlueDot Impact (2026): foundational AI systems, AGI pathways, and societal implications.
Perplexity’s AI Business Fellowship & Agents Intensive with Google — 2025
Procurement and Cybersecurity Trainings — DAI (2023–2024)
Project/Program Management for Development Professionals — Humentum / PM4NGOs (2019, 2022)
CEU SPP: Action Research, Effective Writing, Visualization and Advocacy — 2018
JPO internship — MFA Kazakhstan, Dept. of Multilateral Cooperation — 2016

Selected articles

Article brief: Digital Transformations to Improve the Work and Distribution of the State Scholarships Programs - Kazakhstan scholarship digital transformation, CRM-style workflow, and smart remote management.
Article brief: Data Overload in Fitness Tracking Technologies - Fitness tracking overload, information architecture, and user experience.
Article brief: listening to aid recipients in ‘aid-development’ debate: findings from ‘the Listening Project’ and Jerge-Tal village in Kyrgyzstan - Includes summary, relevance, and source full-text link.

Projects / proof

Case Study: Agenda Intelligence MD — product entry point and evidence-discipline layer
Case Study: Global Think Tank Analyst — horizontal strategic-risk reasoning skill
Case Study: Central Asia + Caspian Hybrid Intelligence Skill — vertical specialist (sanctions, AML, corridors, energy)
Case Study: Gulf + Middle East Hybrid Intelligence Skill — vertical specialist (Iran sanctions, GCC banking, energy, maritime chokepoints)
Case Study: GrantFlow — agent-native grant workflow API (governed agents, deterministic generation, HITL, exports)
Case study: Donor reporting quality and delivery reliability
Case study: Cross-country portfolio delivery and audit readiness
Case study: SaaS / E-Commerce platform launch delivery control
Sample: Activity report
Sample: KPI tracker
Live A2A deployments: Agenda Intelligence A2A (generic strategic-risk triage), Kazakhstan / Middle Corridor Deal Risk Gate (corridor-risk), Kazakhstan Market Entry Readiness Gate (market-entry readiness). Agenstry provider listings.

AI Agent Access

Start here (governance & discovery):

AGENTS.md — canonical rules any AI agent should read first (identity, honesty, evidence, naming, definition of done).
llms.txt — agent-readable site index.
agent-card.json — A2A-style agent card.
agent-discovery.md — discovery guide for Claude Code, Codex, and OpenClaw clients.
agent-match.html — agent-matching guide.

Recruiter-agent routing:

Role snapshots:

Structured profile data:

Profile (Markdown), Resume JSON, Skills JSON, Capabilities JSON, Availability JSON, Recruiter JSON, Engagement Intake JSON.

Verifiable claim & trust artifacts:

Evidence JSON, Verification JSON, Authority JSON.

Observability snapshots (self-reported, see disclaimers in-file):

freshness.json, provenance.json, evals.json.
Eval run: Agenda Intelligence MD, May 2026 — N=3 self-scored protocol eval with full caveats.

Live A2A endpoints (no-payment, public):

Agenda Intelligence A2A — generic strategic-risk triage and source planning: https://agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev/message/send
Kazakhstan / Middle Corridor Deal Risk Gate — region-specific corridor-risk triage: https://middle-corridor-deal-risk-gate-a2a.vassiliy-lakhonin.workers.dev
Kazakhstan Market Entry Readiness Gate — market-entry evidence triage: https://kazakhstan-market-entry-readiness-a2a.vassiliy-lakhonin.workers.dev
Agenstry listings: https://agenstry.com/providers/Vassiliy%20Lakhonin

MCP runtime:

Local server (mcp/server.py) is runnable; remote server live on Cloudflare Workers at https://vassiliy-lakhonin-mcp.vassiliy-lakhonin.workers.dev/mcp (streamable HTTP). Canonical state: mcp-status.json.