Office of the CIO
Houston · Dallas · Chicago
Contract architects, ERP analysts, and a GenAI pod that will not embarrass them in front of the board.
SR SOFT LLC is an E-Verify Company
Trusted advisor · United States · E-Verify
Est. two decades · E-Verify · all visas
We are the trusted advisor to leading organizations across the private and public sectors. Our scale, scope, and knowledge — deep functional and industry expertise, and nationwide reach — let us take on the talent and systems problems that matter most. We work with our clients as we do with our colleagues: we help them build capabilities, deliver measurable impact, and create lasting change.
20+
Years helping clients on the issues that matter
24
Industries. Functional depth. Geographic reach.
US
Headquarters in Texas. Delivery nationwide.
AI
Generative AI at production standard
Featured insights

Healthcare Business Analyst
Cures, information blocking, FHIR R4, USCDI, TEFCA, CMS-0057 — the BA who can read a CapabilityStatement and write the exception legal will sign.
SAP BTP series
SAP practices
Healthcare Business Analyst
SAP BTP series
Featured insights
Wellhead to settlement: DOI, allocation, valuation, royalty, S/4 migration.
Featured insights
OTC, PTP, RTR — the taxed penny, the certificate, the return.
Featured insights
CDM oversight, database lock, GAMP 5, CSA, and GxP AI.
Featured insights
HIPAA, HEDIS, revenue cycle, FHIR, TEFCA, CMS-0057.
Featured insights
RTR, Group Reporting, Central Finance, Concur, CCM.
Featured insights
EWM, TM, Ariba, Public Cloud, Signavio, LeanIX, CALM.
How we help clients
Hire for judgment, coach for context.
We diagnose the problem as it exists on the floor — the system, the statute, the close calendar, the vendor seam — and set a fact base before anyone staffs a slide.
PRA with FICO. Vertex with pricing. CDM with SDTM. Deep functional and industry expertise in one cell — we work with our clients as we do with our colleagues.
One network, one plant, one jurisdiction matrix, one lock path. We prove a slice, then scale. Blueprint follows impact, not the other way around.
Exception owners, recon, hypercare as a designed factory. We leave the organization able to run the work — lasting change, not a visit.
01 — Generative AI
We staff and deliver the unfashionable half of AI: retrieval that cites, agents that are allowed to touch a system of record, evals a second line will sign, and the factory around the model.
Retrieval that cites the procedure, the policy, the claim file — not a vibe. Chunking, hybrid search, rerank, refusal. Built so a plant manager or a nurse manager can trust the sentence.
Tool-calling into ITSM, ERP, EHR-adjacent APIs, and the swamp of internal SOAP. Guards, human-in-the-loop, audit trails. No autonomous chaos on a regulated desk.
Gold sets, rubrics, regression, red-team. The unglamorous work that lets a CISO and a model-risk officer sleep. We staff the people who have done this, and we build the harness when the client does not have one.
Promotion gates, GPU cost, canaries, lineage. Training jobs that rerun without the scientist. The factory around the model — because a notebook is not a product.
PDFs, invoices, inspection photos, call audio. OCR, layout, ASR, multimodal fusion into the same retrieval stack as the text work. Where most of the enterprise still lives.
When prompting is not enough: SFT, LoRA, distillation on the corpus the client actually owns. Pharma literature, claims notes, maintenance logs, legal memos.
02 — The house
AI sits on top. The floor underneath is twenty years of IT staffing.
Enterprise generative AI: RAG, agents, evaluation, MLOps, and multimodal systems — staffed and delivered to production standard.
Contract, contract-to-hire, and direct-hire technical talent — the core we have run for more than twenty years, now including the AI bench.
The desk we have staffed the longest after core IT: ECC to S/4, BTP, cutovers, and the analysts who have actually touched the module.
Enterprise applications, APIs, and platforms — modern stacks that have to live next to SAP, the EHR, or the core.
Dashboards, semantic layers, and reporting the operator will open — Tableau, Power BI, Fabric — plus the data people underneath.
Cloud migration, architecture, and cost optimization across AWS, Azure, and GCP.
Day-to-day IT operations, help desk, and infrastructure support so your team can focus on the business.
AI-driven digital transformation for the same floors we staff: roadmap, vendor choice, and the first production slice — not a slide.
Risk assessments, compliance support, and security hardening to protect your business and your data.
03 — Book of business
We do not paste logo walls. We name the floor: who buys, where they sit, what they actually requisition. That is the book.
Houston · Dallas · Chicago
Contract architects, ERP analysts, and a GenAI pod that will not embarrass them in front of the board.
Gulf Coast corridor
People who have stood in a control room — and can now wire RAG to procedures, permits, and work orders.
Texas · Florida · Midwest
Epic-adjacent analysts, integration engineers, and document AI that respects HIPAA.
NYC metro · Charlotte · remote
MLEs who can talk to validation. LLM evals that a second line of defense will sign.
National
ECC-to-S/4 people, BTP, and ticket-deflection agents trained on the actual queue.
Austin · Bay Area · remote U.S.
Applied AI engineers, MLOps, and multimodal talent hired to the same bar as core engineering.
Nationwide
Surge benches: Java, integration, data, and now LLM engineers on 2–4 week notice.
Texas · Florida
Cleared-enough, patient-enough talent. Modernization without a science project.
Industries we have worked
SCADA-adjacent IT, data historians, turnaround staffing, and RAG over procedures.
Plant IT, LOTO/permit systems, quality labs, and S/4 on the Gulf Coast.
OT/IT, outage systems, GIS, and regulated change control.
Epic-adjacent analysts, integration, and document AI that can sit next to an EHR.
Quality docs, literature RAG, validation-minded engineers, GxP-aware delivery.
Claims platforms, prior-auth, member portals, and model-risked NLP.
Core-adjacent apps, payments, and agents that are allowed read-only.
Policy admin, claims notes, actuarial data, and second-line evals.
Trade support apps, risk data, and people who have sat near a desk.
API platforms, ledger-adjacent work, and cloud that a CISO will sign.
MES, quality, vision on the line, and ERP that matches the floor.
Plant systems, warranty data, and integration across the tier stack.
PLM, quality, and engineers who can work inside a controlled shop.
OSS/BSS, network data, and large-scale Java / integration benches.
Merch systems, order management, and demand models that are not a demo.
WMS/TMS, forecasting, exception queues, and planner copilots.
Fleet, dispatch, and the unglamorous systems that keep freight moving.
Modernization without a science project. Patient staffing. Texas and Florida.
SIS/ERP, research computing, and IT that has to survive a semester.
ECC, S/4, BTP, cutovers, AMS — the desk after core staffing.
Product engineers, platform, and the first ML hire who will own on-call.
Surge benches for SIs: Java, data, SAP, and now LLM engineers.
Content platforms, rights data, and cloud that has to be cheap at scale.
PMS-adjacent, loyalty, and the integration swamp behind a reservation.
Selected clients
Leading companies across life sciences, energy, industrials, travel, financial services, and the public sector. We help them staff and deliver the systems that cannot miss.
04 — Representative work
Gulf Coast energy
A refiners’ turnaround library — permits, LOTO, OEM manuals — wired into a grounded assistant. Staffed a GenAI engineer and an MLOps lead onto a plant IT desk that had never hosted a model. The win was not the demo. It was operators using it on night shift.
Regional health system
Prior-auth packets and referral PDFs. HIPAA-aware pipeline, human review, no clever leakage. We placed the informatics analyst and the applied NLP engineer as one pod so clinical ops did not have to translate.
National insurer
Summarization and coding assist over a decade of notes. The staffing brief was “someone who has been yelled at by model risk.” We sent that person. Then the eval harness.
SAP competence center
Years of incident text, not a generic ITSM bot. An ERP analyst who knew the module sat with an LLM engineer. Deflection went up because the answers sounded like the shop, not like a vendor brochure.
Discrete manufacturing
Defect images plus the quality SOP. Multimodal retrieval for the quality engineer; a smaller model at the edge where the plant would not send frames out. Contract-to-hire on both seats.
Commercial bank
Read-only tools, dual control, every call logged. The architect we placed had shipped payments software before they shipped agents. That order matters.
Pharma / life sciences
A scientific RAG stack with citation discipline a QA group would accept. Staffed an applied scientist and a document engineer. Sponsorship in flight before the start date.
3PL / logistics
Classical demand models first — because GenAI does not fix a bad number. Then a planner-facing assistant over the forecast and the exception queue. MLE plus a full-stack engineer, same vendor, one throat to choke.
05 — Platforms
Not a logo-rental wall. These are the platforms we staff, implement, and keep running — the same set a Houston enterprise shop actually buys.
Cloud, data, Bedrock / SageMaker benches
Azure, Fabric, OpenAI Service, Dynamics
ECC, S/4HANA, BTP, integration
Cloud apps, databases, middleware
CRM, Service, MuleSoft-adjacent
Dashboards the floor will actually open
Locations
Headquarters · Texas
18722 University Blvd Suite 200
Energy · SAP · plant IT
Gulf Coast delivery
Nationwide bench
Remote and on-site
Insights brief
Next requisition