Intake and follow-up systems
Classify emails, calls, forms, and customer threads. Draft the reply, assign an owner, surface anything that's gone cold.
Charleston, SC · AI automation consulting · custom ops systems
Practical automation for the work your business keeps doing by hand.
I map your workflows, find the highest-value places to automate, then build custom systems and dashboards that save time without handing risky decisions to AI.
About
Background
I run Blaine Heffron Consulting. I help small businesses find the parts of their workflow that are wasting time: messy spreadsheets, missed follow-ups, slow reports, scattered customer or inventory data. Then I build the system that keeps that work organized so you can spend more time talking to customers and improving the business.
I have a PhD in physics from the University of Tennessee. I worked on the PROSPECT neutrino experiment at the High Flux Isotope Reactor at Oak Ridge National Lab. My dissertation focused on understanding background radiation produced by the reactor and using machine learning to improve background rejection.
Recent work: a private automation dashboard platform, StockForge, Songcraft, Stellar tooling, and custom software for teams stuck in process.
I'm most useful when the problem involves messy workflows where you are unsure about what AI can and cannot automate.
Business systems automation
Internal tools
AI systems
Integrations
Reporting pipelines
Machine learning
Scientific computing
Signal processing
Detector analysis
TypeScript
Python
Rust
Node.js
React / Preact
Automation design
Technical product development
Research software
Consulting
Business automation systems
AI automation, internal tools, ops dashboards, and technical software for teams whose work is spread across too many places.
The starting point usually shows up in the numbers. Product data that will not reconcile. Orders that need a status board. Follow-up that depends on memory. Hours spent shuffling data nobody asked for.
Pilot availability
A build that has to earn its keep
I'm trying a performance-based structure with a few businesses. Pick one bottleneck. Set the baseline. Build the smallest useful version. Upside only kicks in when the number moves.
- Fits: SKU cleanup, order visibility, slow quoting, missed follow-up, inventory mistakes, repetitive docs, admin work that's eating billable hours.
- Doesn't fit when the numbers are vague or nobody owns the workflow.
- We agree on attribution, support, payment triggers, and what happens if the metric doesn't move — before anything gets built.
Quote and proposal workflows
Messy inputs in, reviewed quotes out. Approval queues people will actually use, with status the owner can see.
Inventory and order cleanup
Reconcile product data and SKU mappings across tools. Surface shipping exceptions, stuck orders, and handoffs that fall through the cracks.
Owner-ready ops dashboards
One private view for the work that needs action this week: process status, approvals, exceptions, connector health. Skip the vanity charts.
Lane 1
Automation and ops cleanup
For founder-led companies where manual systems keep creating the same status, data, or handoff problems.
Business process audit
How the work moves now, where it jams, and which system should be automated first. Usually the answer surprises someone.
Fixed diagnostic, often credited into an approved build
Start here if the work feels clunky and nobody can point at the right first step.
Audit detailsMeasurable automation pilot
One painful business system, one short build. Usually intake, reporting, inventory/order visibility, CRM cleanup, or document handling.
Lower upfront build cost, with upside only when the agreed metric moves
Pick this if you already know the bottleneck and can attach a number to it.
Pilot detailsCustom ops cockpit
One private dashboard pointed at the work that needs attention: stale tasks, pending approvals, exceptions, unpaid invoices.
Scoped dashboard build with connectors and handoff
Pick this when the work lives in too many tools and the owner has lost the operating view.
Dashboard details
Lane 2
Scientific ML and technical analysis
For research and engineering groups whose problem isn't a no-code one. ML, analysis tooling, software where the technical calls matter.
Technical advisory
Consulting on analysis strategy, system design, or a debugging pass before someone signs off on a bigger build.
Scoped advisory, milestone, or retainer structure
Use this when the project deserves a real technical pass before anyone starts shipping.
Technical detailsCustom technical buildout
ML pipelines, detector and signal analysis tooling, research software, technical dashboards. The work where plausible isn't good enough.
Scoped after discovery with milestones and handoff defined up front
Use this when the work needs real software and somebody has to make the technical calls.
Technical detailsNot sure which lane fits?
If the headache is reporting, approvals, inventory, quoting, fulfillment, or customer follow-up, start in the automation lane. If it's ML, simulation, or research software, the technical lane is the right door.
Proof
Selected work
A handful of recent projects across ops software, developer tooling, and technical systems.
Automation dashboard platform
My private automation engine for client dashboards: workflow queues, approval gates, connector health, agent-assisted drafts, and auditable operating views.
StockForge
Inventory and operations software for a chemical-industry use case: BOMs, barcode scanning, reorder logic, and supply-chain workflows in one place.
Songcraft
A music reading game where kids play along with sheet music in real time, with classroom tools, parent tracking, and live note matching over WebSockets.
Breeze Homes
A cloud-backed platform for customer onboarding, project tracking, floor plans, photo storage, and contractor workflows.
Scaffold Stellar
Developer tooling for building and shipping Stellar smart contracts with less setup friction.
EquitX
Contract and backend support systems for CDP indexing, price feeds, liquidity tracking, and liquidation logic on Stellar.
soroban-sdk-tools
Rust proc macros and test utilities for Soroban contracts, funded through an SCF #39 Developer Tooling grant.
paper-search-mcp
MCP server for searching and indexing scientific papers across multiple sources, backed by local hybrid search.
symbolic-math-mcp
MCP server for symbolic mathematics and tensor algebra in specialized technical work.
ML for PROSPECT
CNN and GCN-based event reconstruction for the PROSPECT neutrino detector, improving usable signal statistics by 3.3%.
Contact
Get in touch
If a workflow is wasting time, leaking margin, or stuck in admin land, send me a note.
A good first message tells me what's broken, which tools touch it, how you measure the problem today, and roughly when you want it gone.
Research
Selected publications
Academic work in neutrino physics, detector analysis, and machine learning.
-
JINST 20, P08006 (2025)
-
Final Search for Short-Baseline Neutrino Oscillations with the PROSPECT-I Detector at HFIRPreprint (2024)
-
Reactor Antineutrino Directionality Measurement with the PROSPECT-I DetectorPreprint (2024)
-
Phys. Rev. Lett. 131, 021802 (2023)
-
Joint Measurement of the 235U Antineutrino Spectrum by PROSPECT and STEREOPhys. Rev. Lett. 128, 081802 (2022)