A human-centric AI ecosystem

Orchestrating a human-centric life.

AVAM is Mike Oh's evolving personal AI ecosystem—designed to connect life management, professional workflows, infrastructure, and community initiatives through two purpose-specific AI orchestrators.

Explore the Ecosystem
Discover
By the numbers

AVAM at a Glance

Verified facts from the live system, audited September 2026.

24
Scheduled automation jobs running continuously
13+
Infrastructure checks performed every 5 minutes
5
Specialist reasoning personas (Atom, Pax, Morgan, Sterling, Vida)
Cloud
+ VPS
Managed cloud hosting with a self-operated observability server
Zero
Analytics, tracking cookies, or monitoring scripts on this website
The AVAM philosophy

More than AI. A foundation of trust.

AVAM is built on a simple, powerful premise: technology must serve humanity with unwavering integrity. Our name reflects our family values—the core of our commitment to data privacy, ethical responsibility, and purposeful innovation. This is not just a project; it is a personal mission to build a future where AI augments our lives with respect and transparency.

Portrait of Mike Oh, founder and architect of AVAM
"We are not building a tool. We are cultivating a trusted partner for life's journey, grounded in principles that matter."Mike Oh · Founder and architect of AVAM30 years IT · RFID patents · licensed PI · Hudson coach →
Two orchestrators, one purpose

A dual-core ecosystem

AVAM brings together a personal orchestrator and a dedicated community AI—each distinct, deeply aligned.

Larry logo

Larry: Your Personal AI Orchestrator.

Larry is the central intelligence of the AVAM ecosystem—a second mind, executive assistant, and infrastructure manager connecting health, work, and daily life.

Larry runs 24 scheduled automations daily: morning briefings, health monitoring, email triage, finance checks, infrastructure self-healing, and weekly intelligence reports.

01

Daily Intelligence

Morning briefings, JIT meeting prep, and weekly reviews synthesised from real data.

02

Health & Wellness

Biometric monitoring with threshold alerts and Garmin data integration.

03

Infrastructure

Self-healing monitor with 13+ checks every 5 minutes and automatic rollback.

Explore Larry →

Max logo

Max: AI for Community Impact. Live

Max is Mike's live advisory platform—supporting structured mentoring relationships, career journey tracking, and community impact work with confidentiality and human oversight at its core.

Max is a deployed advisory platform with session management, career journey tracking, an AI chat assistant, and Google Workspace integration—purpose-built for structured one-on-one mentoring.

01

Structured Advisee Support

Career journeys, session notes, action items, and application pipeline tracked end-to-end for each advisee.

02

AI Advisory Assistant

Max Chat draws on each advisee's profile and session history to support in-the-moment advisory decisions.

03

Confidential by Design

Account isolation, access-controlled database, and human review on all outgoing communications.

Learn About Max →

Verified live workflows

A Day with Larry

Every entry below is a confirmed, deployed automation—verified from source code in September 2026.

Mon 5:00 AM SGT
Memory Hygiene Larry
Orphaned records purged from the vector memory and knowledge index after any deletions from the prior week.
Daily 6:00 AM SGT
Finance Check Sterling
Programmatic scan of portfolio thresholds using live stock data. Alert sent to private messaging if a threshold is breached; silence means all clear.
Daily 7:00 AM SGT
System Digest Larry
Infrastructure health snapshot—disk, memory, and cloud deployment status—sent to private messaging each morning.
Daily 7:30 AM SGT
Email Triage Pax
Inbox scanned in parallel; Action Required messages become Gmail drafts for Mike to review and send. FYI and Archive items are skipped.
Mon 8:00 AM SGT
Weekly Tech Intelligence Pax
Three AI research queries (AI, Enterprise SaaS, Cybersecurity) synthesised into an HTML report delivered to inbox and private messaging.
Daily 8:00 AM SGT
Morning Briefing Larry
Calendar, priorities, and context synthesised into a focused daily brief delivered via private messaging.
Daily 8:30 AM SGT
Smoke Test Larry
Five key endpoints hit and latency measured. Results logged and sent to private messaging.
Daily 9:00 AM SGT
Health Check Vida
Biometric thresholds checked against the health database. Alert sent only when a reading warrants attention; silence means all clear.
Mon 10:00 AM SGT
Agentic Research Pax
AI-driven research synthesis on topics of interest, delivered to inbox via AI research and web search.
Daily 10:00 AM SGT
Finance Update Sterling
Financial context synthesised and delivered as a private messaging update or inbox briefing.
Mon (weekly) & ongoing
Industry Roundup Morgan
Competitive intelligence for professional context, synthesised from industry sources and delivered to inbox.
Sunday 6:00 PM SGT
Weekly Review Larry
Journal, tasks, health, and finance records from the week synthesised into a structured review delivered to inbox.
Daily 9:00 PM SGT
Wellness Check Vida
Evening health threshold review. Programmatic check only—alert sent to private messaging if a metric needs attention.
Daily 11:30 PM SGT
Episodic Memory Larry
The day's conversation summarised and archived to the episodes table and vector memory for long-term context and recall.
Every 5 minutes
Self-Heal Larry
13+ infrastructure checks: database connectivity, schema integrity, OAuth status, cloud deployment health, log scan, task backlog, disk, memory, private server liveness, AI model quota, and more. Auto-fixes schema drift; auto-rolls back if repeated failures detected.
Every 10 minutes
Flight Tracker Larry
Real-time flight status queried when a flight is active. Update sent to private messaging.
Every 15 minutes
Meeting Prep Larry
Calendar checked for upcoming appointments. JIT briefing alert sent to private messaging if a meeting is within the window.
Saturday (weekly)
Code Audit Larry
Automated SAST and security review of the codebase. Report generated and delivered.

All times Singapore Time (SGT). Event-driven automations (self-heal, flight tracker, meeting prep) fire on trigger, not clock schedule. Verified from automation job configuration and Flask route definitions, September 2026.

System architecture

Under the Hood

A verified map of how AVAM's components connect. Only confirmed, deployed components are shown.

Mike Oh
Web UI · Messaging
↓
Personal Domain — Live
Larry — Python / Flask
Managed cloud · larry.avam.buzz
↙
↓
↘
Automation Layer
24 scheduled jobs
Enterprise AI
Reasoning & synthesis
Data Stores
Relational DB · Vector DB
↓
↓
External APIs
Calendar · Health
Search · Research · Flight
Private VPS
AI call observability
Webhook handler
· · · Community domain boundary · · ·
Community Domain — Live
Max — Advisory Platform
Deployed · Railway · separate repository

Internal IP addresses, credentials, and webhook secrets are not shown.

Work with Mike

Engagement Paths

AVAM is an evolving personal AI ecosystem and advisory blueprint—not an off-the-shelf software product.

Is this for you?
  • You want a private AI system built around your life or organisation — not another SaaS subscription to evaluate and cancel.
  • You need someone who has actually built and operated this kind of system, not consulted on AI strategy from the outside.
  • You value owning the result: a working architecture and blueprint you control, not a managed service or vendor dependency.
Most requested

AI Architecture Advisory & Blueprint

How to design and operate a private AI ecosystem—from principles to infrastructure to daily workflows. Includes a custom architecture blueprint tailored to your personal or organisational context and ethical constraints. Typically a private 1:1 engagement over 2–4 weeks.

Get the Advisory Blueprint →Private 1:1 · remote · 2–4 weeks

Workflow & Automation Design

From trigger to output: end-to-end automation design that prioritises human oversight and auditability.

Enquire →

Speaking & Demonstrations

Practical AI orchestration for real life and professional work—grounded in evidence, not aspiration.

Book a session →

Community-Impact Collaborations

Selected partnerships aligned with AVAM's values of dignity, confidentiality, and meaningful contribution.

Explore →

Custom Private-AI Blueprint

Architecture documentation tailored to your personal or organisational context and ethical constraints.

Enquire →

Not sure which fits? See all paths

A working ecosystem

Evidence over aspiration.

Larry operates as a live system—running scheduled automations, processing real data, and delivering daily outputs. Max is a live advisory platform supporting structured mentoring and community impact work. Together, they represent a working approach to human-centric AI orchestration grounded in transparency: every capability on this site is classified as Live, Planned, or Illustrative.

Read Our Principles
Proof of work

Built, not claimed.

AVAM's credibility comes from its build record. Every milestone below is verified from source code and system logs—the same standard applied to every claim on this site.

2024

Larry v1 — first working orchestrator

Morning briefings, email triage, biometric monitoring, and infrastructure checks deployed and running daily. Single-user, private-by-design from day one.

2025

Automation scale-up & five specialists

Grew from 6 to 24 scheduled jobs. Atom, Pax, Sterling, Vida, and Morgan added as specialist reasoning personas. Self-healing monitor with auto-rollback deployed.

2025

Max — advisory platform launched

Community advisory platform deployed on Railway: session management, advisee journey tracking, AI assistant, and Google Workspace integration — purpose-built for structured 1:1 mentoring.

Sep 2026

Full system audit

All 24 automations verified from source code. TOTP MFA confirmed live (v2.2.6). Private VPS observability layer operational. 959 vector memory nodes indexed.

No advisee outcomes are cited here. Max handles confidential mentoring relationships — that data stays private by design. The build record is the proof.

Verified controls

Privacy in Practice

Controls listed here are verified from implementation—not stated aspirations.

ControlHow it worksScopeStatus
Data separationPersonal and community data handled in separate contexts; Larry uses a context column (personal / official) on key tablesLarry (live); Max when deployedLive (Larry)
Human approvalHigh-impact actions gated in code; email drafts require Mike to review and send before transmissionAll outgoing communicationsLive
Access limitationNo public accounts; single-user system with TOTP multi-factor authentication (added v2.2.6)Entire Larry ecosystemLive
No website trackingNo analytics, session cookies, or tracking scripts on this website—verified by code inspectionPublic website (avam.buzz)Live
Conversion signal loggingOn explicit form submit or primary CTA click: enquiry path and type logged server-side to Mike's private infrastructure. No names, emails, or message content in any log. Browsing and page views generate no signal.Advisory, Speaking, Partnerships forms; primary engagement CTAsLive
Observability loggingAI call metadata logged on a self-managed private server operated by Mike Oh; not shared with third partiesLarry AI callsLive
API retentionExternal AI and research providers govern their own server-side retention per their published policiesExternal API callsPer-provider