Alex runs Om Concepts as the only principal. Years of consulting, training, building, and implementing on AI, agents, sales, and business operations for clients, now consolidated into one founder-led practice.
Alex runs Om Concepts as the only principal. He started college in biomedical engineering, and the lab habits stuck: measure before you move, diagnose before you prescribe, keep a record of what you changed. Those habits show up in every engagement.
Eight years inside a Fortune 500 sales organization taught him how enterprises actually execute and where they lose. Since going independent he has spent years consulting, training, building, and implementing for clients across AI, agents, sales, and business operations. Om Concepts is where that work consolidates under one practice.
He works across frontier models from Anthropic, OpenAI, Google, Meta, xAI, and Perplexity, alongside open-source weights and local inference. Every comparison comes from direct use. He is deep on agent harnesses, MCP, multi-agent orchestration, and every modality that matters: voice, image, video, audio, and code.
Delivery is founder-led with tight scopes, usually a 30 to 90 day window with a defined outcome. Remote by default, in person when a local engagement benefits from it. The pitch is simple: the Lehigh Valley has plenty of agencies, and almost none of them can actually ship modern AI systems. Om Concepts fills that gap.
“The Lehigh Valley has plenty of agencies. Almost none of them can actually ship modern AI systems. That's the gap we fill.”
Why Om Concepts exists as a local shop
“The person on your sales call is the person writing the code. One engagement, one principal, weekly check-ins.”
To a prospect asking who actually does the work
“When you're burning credits, dude, it's like gambling. We cache, batch, run open-source where it fits, and show you the bill.”
On how agents should handle money
Credentials
What Alex brings to the table.
Operator experience, engineering habits, and hands-on AI work. Each line is a shipped receipt somewhere on this site or in client engagements.
01
Fortune 500 operator
Eight years inside a Fortune 500 sales organization.
02
Engineering roots
Started college in biomedical engineering. The lab habits stuck.
03
Solo to Fortune 10
Client work spans one-person operators up to Fortune 10.
04
Years on AI delivery
Consulting, training, and shipping AI and ops systems before it was fashionable.
05
Every model tier
Frontier-lab models, open-source weights, and local inference in daily rotation.
06
Agents in production
Custom harness, MCP servers, and multi-agent handoffs.
AI experience
Modalities, agents, MCP, orchestration, and lab programs. Read the full breakdown.
The path that led here. Engineering school, eight years inside a Fortune 500, multiple businesses operated, and now an AI-first agency.
Early foundations
Biomedical engineering and enterprise sales
Started college in biomedical engineering. Carried the lab habits forward: measure before you move, instrument everything, keep the protocol tight.
Fortune 500 operator
Sales and operations at scale
Eight years running the full sales motion inside a public company. Saw the systems behind enterprise selling, the pace of operations, and where most enterprises lose execution.
Founder operator
Multiple businesses across services and software
Started, built, and operated companies across marketing, software, and AI. Direct exposure to every layer of running a business, from intake to payroll to shipped product.
Om Concepts, now
AI-first agency, founder-led
An agency built around agents from day one. Founder-led delivery for small and mid-sized businesses in the Lehigh Valley and nationwide by remote.
How agents are built
Model, harness, environment.
Every agent on every client engagement composes the same three primitives, inside the same three guarantees. The full toolbox lives below.
Stack
A running snapshot of the stack.
A sample across ecosystems. Updated weekly.
Agent harnesses
What we run
Claude Code
Codex
Cursor
Open Claw
Manus
GenSpark
Pplx Computer
Nemo Claw
Frontier model providers
Models we work across
Anthropic
OpenAI
Gemini
xAI
Meta
Pplx
Open weights
Weights we benchmark
GLM
Kimi
Qwen
Gemma
Nemotron
DeepSeek
Mistral
Local inference
Where models run on our own hardware
Ollama
LM Studio
llama.cpp
vLLM
MLX
HF
MCP ecosystem
Tool surfaces we expose to agents
MCP
Notion MCP
context7
Chrome DT
Fireflies
Workspace
Cloudflare MCP
Research and knowledge
How we stay current
Pplx Pro
context7
arXiv
DeepWiki
HF papers
Developer environment
What we ship from
Git
GitHub
Node.js
Docker
Raycast
Warp
Obsidian
1Password
Agent orchestration
How tool-use loops get built
Anthropic SDK
OpenAI SDK
MCP SDK
Client ops and CRM
Client-side stack
Notion
Linear
Airtable
Zoho
Fireflies
Workspace
Communications
How we reach people
Resend
Slack
Gmail
WhatsApp
Ads and analytics
Spend and measurement
Google Ads
Meta Ads
LinkedIn Ads
GA4
GTM
Metricool
Cloud and infra
Where it ships
Vercel
Supabase
Cloudflare
GH Actions
Commerce and web
Client stores and sites
Shopify
WordPress
Squarespace
Stripe
Plus a constantly shifting set of CLIs, MCPs, and personal tools that never make a public list. If it helps a client ship, it gets considered.
How an engagement runs
Five phases. One shipped system.
Where we work
Lehigh Valley first. Remote everywhere else.
We are based in the Lehigh Valley and work in person across Allentown, Bethlehem, and Easton. Remote engagements run nationwide and into the EU when the timezone fit allows.
Positioning
The Lehigh Valley's first AI-first agency
Most agencies bolted AI onto an existing service menu. We built the practice around agents from day one. Agents are the operating system of how we research, strategize, build, and measure.
That means different deliverables: visible agent workflows, receipts for every AI-produced artifact, and systems you can keep running without us.
How we work
Founder-led. No layers. Real artifacts.
You work with the person doing the work. No account manager hand-offs. No deck-only engagements.
Every claim on this site is tied to a shipped artifact or a documented outcome. When an agent produces something for you, you get the receipt: model, prompt hash, token counts, latency.
Engagements are scoped tightly. We prefer 30 to 90 day windows with a clear deliverable over open-ended retainers.
What we build
Marketing, software, and agents in the same engagement
The value comes from the overlap. A brand system that the agents author against. A content pipeline that the SEO strategy compounds into. Internal tools that actually match how you work.
See the services page for current offerings and the proof page for shipped examples.
Why Lehigh Valley
Local presence, remote execution
We're rooted in the Lehigh Valley because the market here is underserved by agencies that can actually ship modern systems. We serve clients nationwide remotely, and we show up in person for local partners.
