Software & Product Engineering
ShieldStone Labs
Austin, Texas The software your firm runs on Currently building for law firms

You run a law firm.
We build the software it runs on.

The operational software a practice actually depends on. Intake, case management, and the systems that quietly eat your team's hours.

§ 01

Capabilities

Most firms arrive with a problem, not a product category. Part of the work is telling them which of these it actually is.

Product engineering

Custom software shaped around how your business actually works, not forced into a tool built for someone else.

Systems & automation

Your tools don't talk to each other. We connect them, so your team stops re-typing the same information across systems.

AI engineering

AI where it earns its place, and a straight answer when it doesn't.

§ 02

Case studies

Regulated industries, sensitive data, real consequences for error. The work a firm needs, proven on businesses that look a great deal like one.
Case Study 01 · Finance
Intertwined Investors logoIntertwined Investors

An automated trading system for an investment firm.

Based inPortland, OR
SectorInvestment / Trading
Asset classStocks & ETFs
Strategies in parallelUp to 50
In production4 months
StatusOngoing

The firm had spent roughly six months on an earlier build, left unfinished and structurally fragile, with real capital intended to run through it. After a full review, ShieldStone recommended a clean rebuild, on the record, before the engagement began.

The system ingests live market data, runs the firm's signal algorithms, scores each signal for probability of success, and manages position opens and closes without continuous human involvement. Every signal and every trade is logged with the exact market conditions it fired in. A testing framework runs up to fifty strategy variations at once, so refinements are validated against each other before a dollar is committed.

Case Study 02 · Logistics

A production platform a waste-hauling operator runs its business on.

Based inAustin, TX
SectorLogistics / Operations
Revenue run through it$1M+ / yr
Dumpsters managed130
DatabasePostgres RLS
StatusOngoing

The product owner came with a working MVP that would not scale and a security model too thin for an international rollout. The choice was deliberate: patch toward a target the backend could not reach, or rebuild to the level the rollout required. The rebuild was the proportional answer.

The platform now handles scheduling, dispatch, billing, and asset tracking. A photo-and-GPS workflow gives dispatch an exact pin for every dumpster instead of a written description, and analytics surface which assets and customers actually carry the margin, so the fleet grows on data rather than guesswork.

Its first operator is on track to run over a million dollars of revenue through the platform this year.

Case Study 03 · Artificial Intelligence

Benchmarking and prompt engineering for an AI startup.

Based inAustin, TX
SectorAI / Product
BuiltEvaluation layer
ResultAccuracy ↑
StatusDelivered, in use

The product relied on AI output with no way to measure whether it was getting better or worse. ShieldStone built the evaluation and integration layer the team lacked, then refined the prompts against it to raise output accuracy measurably.

The engagement was delivered and handed off. The system remains in use today.

Case Study 04 · Insurance

The day a brokerage lost all email, and got it back.

Based inPortland, OR
SectorInsurance
IssueFirm-wide email outage
ResolutionSame day
StatusDelivered

The entire firm could not send or receive email. For a brokerage that runs on client correspondence, every hour offline is business not getting done. ShieldStone diagnosed the Outlook failure, traced the cause, and restored email across the firm the same day.

Not every engagement is a ground-up build. Some are about being who a business can reach when something critical breaks and the clock is running. This was that.

§ 03

Method

What most firms selling software will not put in writing.

"We tell you when you don't need AI, and build the right thing instead."

i.

The right tool, honestly

If a problem doesn't call for AI, we say so. You get the solution that fits, not the one that sells.

ii.

You own what matters

Your cloud, payments, and data are in your name from day one. Code terms are set explicitly up front, transferred or licensed, never a black box you can't leave.

iii.

You pay on proof

Payment is tied to milestones you see working in front of you, not progress reports and promises.

§ 04

How we work

What an engagement looks like from your side of the table.
01

Diagnose

It starts with a conversation, no charge. If it's a fit, a focused paid discovery phase maps the real problem and returns a concrete plan and quote, yours to keep whether or not we build it, and credited toward the build if we move forward. Sometimes the right answer is smaller than you expected.

02

Scope in phases

A plan broken into fixed milestones, so a long build is taken in stages rather than carried as one large risk up front.

03

Build on proof

You see each milestone working before it is paid for. Demonstrated, not reported.

04

No lock-in

Accounts in your name, data that's yours, code terms agreed up front. Nothing here traps you with us.

§ 05

Selected work

Breadth behind the build.
Legal

Designing a productized client intake and case-management platform for an Austin immigration firm: guided intake, a secure client portal, and a staff workspace for the whole matter lifecycle.

Legal

Scoped an AI-driven intake screening and sales support system for an Austin immigration firm: case triage against the firm's own criteria, and on-demand support for the intake team.

Web

Designed and built the website and lead funnel for Hidden Gem Gym, an Austin boutique studio. Over $120,000 in revenue generated through the site in its first year.

§ 06

Who you work with

Ennis M. Salam

I started ShieldStone Labs because too many businesses get sold AI they don't need and software they can't own.

I'm an engineer first. People come to me for AI, and often, once we talk the problem through, what they actually need is something quieter: automation, or a system built properly, or the right tool instead of the fashionable one. I'll tell you which it is, plainly. Your accounts and your data stay yours, the terms for the code are set up front, and nothing locks you in.

Most of this work has come by referral, one client introducing the next. When we work together, you deal with the people building your system, not a sales layer sitting in front of them.

Ennis M. Salam, Principal · ShieldStone Labs
University of Texas at Austin · Electrical & Computer Engineering
Cisco · CX Cloud
Scale AI · Led a 23-engineer team
LinkedIn ↗
Get in touch

Tell us what your firm runs on.

No pitch. Tell us where your team loses the most time, and we'll tell you straight whether we can help.

Start a conversation

Austin · Dallas · San Antonio · Portland · Las Vegas · and wherever you are

ShieldStone Labs Software & Product Engineering · Austin, TX © 2026