Company — HyperLinxWork

OUR STORY

From LinxWork
to HyperLinxWork.

LinxWork was founded in 2008 as a network systems integrator and software house in Bangkok. From the start we worked with carriers that operate networks for an entire country — relationships that compound, not rotate.

Three anchor customers have shaped how we build software for the network. We began working with a tier-1 Thai telco and the Laos national telco in 2010, and with a Laos government ministry in 2015. Each relationship is still active today — 11 to 16 years each.

Across those relationships we shipped real infrastructure work: WiFi metro, core/backbone integration, firewalls, AAA, mail gateways, load balancers, branch-router rollouts, multi-ISP load balancing, financial e-service infrastructure, IPAM, central configuration management. In 2019 we deployed AutoProvision at the tier-1 Thai telco — our network automation product, now in its third phase.

In 2022 we made a deliberate choice: AI was about to change network operations, and we didn't want to wrap someone else's foundation model. We began building our own AI stack — a nano language model architecture paired with graph neural networks, purpose-built for the reasoning patterns that network operations actually require. Four years later, that stack runs production agents on commodity CPU. No GPU dependency. No cloud-only lock-in.

In 2026 we rebranded to HyperLinxWork to reflect the broader platform: AutoLinx (the agent for network operations, evolution of AutoProvision), SecureLinx (the security agent, in early access for Q3 2026), and FabricLinx (the intelligent traffic layer, for environments that need wire-rate ingest).

Same team. Same Bangkok HQ. Same operators measuring us on whether their MTTR drops next quarter. The thesis has just become sharper.

2008

Founded as LinxWork in Bangkok · operating in the same hands since

3 × 11-16 yrs

Three anchor customers: Thai telco, Laos national telco, Laos government ministry

2022

Began building our own AI stack · nano LM + GNN · CPU-deployable

Products on the platform · AutoLinx · SecureLinx · FabricLinx

WHAT WE BELIEVE

Three convictions guide every decision.

Stated plainly — these are the choices we make repeatedly, across product, hiring, customer conversations.

CONVICTION 01

Reasoning needs full-fidelity input.

Sampling, aggregating, and bucketing are admissions of defeat. We engineer the data path to deliver every packet, every flow, every event — and design the agent to consume it. If the input is incomplete, the reasoning is theatre.

CONVICTION 02

Engineers decide. Agents propose.

We don't believe in autonomous AI for production networks or security. Agents bring evidence and proposals; engineers and analysts decide. Auto-approval is opt-in per action class — never default. The day we trust agents to act alone is not this year, or the next.

CONVICTION 03

Evidence outlives the moment.

Every decision lands in a signed, replayable, vendor-neutral ledger. Six months from now, when the auditor asks why — the answer is there. When the engineer asks "what did we change at 3 AM" — the answer is there. Memory is a feature, not an afterthought.

LEADERSHIP

The team running this.

Today's leadership joined as LinxWork transformed into an AI-native company. Backgrounds in network operations, ML systems, and FPGA — chosen for the work ahead, not the work behind.

CEO · OPERATIONS

[CEO Name]

Chief Executive Officer

Network operations background at telco scale. Led the company through the 2022 AI transformation — repositioning a 14-year-old SI business as a product company, while keeping AutoProvision in production for tier-1 customers.

CTO · AI & SOFTWARE

[CTO Name]

Chief Technology Officer

ML systems background. Designed the nano-LM + GNN reasoning architecture from scratch — choosing CPU-deployability and reasoning-fidelity over scale-of-parameters. Leads the AI research team.

CHIEF ENGINEER · HARDWARE

[Chief Engineer Name]

Chief Engineer

FPGA and packet-processing background. Designed FabricLinx to bring wire-rate observability to the mid-market — using affordable FPGA platforms like the Kria KR260, not just carrier-class Alveo.

The full team is now 18 people across engineering, product, and customer success — actively hiring. See open roles ↓

TIMELINE

Where we've been. Where we're going.

2008

LinxWork founded in Bangkok

Started as a network systems integrator and software house — focused on telco-grade network engineering.

2010

First two anchor customers · Thailand & Laos

Began working with a tier-1 Thai telco and the Laos national telco. Both relationships are still active today — 16 years and counting.

2013 – 2018

Telco-grade infrastructure work

Core/backbone integration, firewall & DDoS, web caching, mail gateway, load balancers, central AAA (SignONE), MAC authentication, centralized config management. The depth that AI alone can't substitute for.

2015

Third anchor customer · Laos government ministry

Network security infrastructure deployment for a Laos government ministry. The start of an 11-year financial-infrastructure relationship.

2019

AutoProvision Phase #1 · tier-1 Thai telco

Our network automation product enters production. Two more phases follow (2020, 2026) — the third with our AI reasoning layer in the loop.

2022

AI transformation begins

The leadership team made a deliberate choice: build our own AI stack rather than wrap someone else's foundation model. Nano language model + graph neural network research begins.

2023 – 2024

Government rollouts & automation products

216-branch deployment for the Laos government ministry. ONT Management System and BPI (Broadband Provision Intelligent) for the Laos national telco. The bridge from manual ops to AI-augmented.

2025

FabricLinx prototype validated

First wire-rate ingest pipeline running in customer environment on Kria KR260 — proving the mid-market FPGA path. Decision to make FabricLinx a product, not a feature.

2026 · NOW

Rebrand to HyperLinxWork · three products · AutoProvision Phase #3

Platform rebrand reflects the broader thesis. AutoProvision Phase #3 ships with our AI reasoning layer in production at the tier-1 Thai telco. AutoIPAM ships at the Laos national telco. SecureLinx in early-access. FabricLinx in select deployments.

2026 · Q3 (next)

SecureLinx early access

Seven design-partner slots open; four taken. GA targeted late 2026.

2027

FabricLinx API · multi-tenant design partners

FabricLinx API exposed for customers running their own SIEM, SOAR, or in-house analytics. Multi-tenant fabric capability opens for select design partners — MSPs, telco wholesale, datacenter operators.

2026 · Q3 (next)

SecureLinx early access

Seven design-partner slots open; four taken. GA targeted late 2026.

2027

FabricLinx API · multi-tenant design partners

CUSTOMERS

Who runs HyperLinxWork in production.

We don't publish customer logos. Our customers run regulated infrastructure where vendor disclosure is sensitive — and we respect that boundary. What we can show: the shape and depth of each relationship.

ANCHOR 01 · 16 YEARS

Tier-1 Thai telco

since 2010 · still active

Relationship spans 16 years. AutoProvision Phase #1 deployed in 2019 · Phase #2 in 2020 · Phase #3 with our AI reasoning layer ships in 2026. Also: VPN speed test, 5G VLL cutover, node integration.

Reference call under NDA

ANCHOR 02 · 16 YEARS

Laos national telco

since 2010 · still active

WiFi metro, core/backbone integration, firewall + DDoS, mail gateway, load balancer, SignONE (centralized AAA), MAC authentication, IPAM, ONT management, BPI, AutoIPAM (2026).

Reference call under NDA

ANCHOR 03 · 11 YEARS

Laos government ministry

since 2015 · still active

Network security infrastructure, data center improvement (authentication, DNS, financial e-service, email, core), 216-branch deployment with multi-ISP load balancing, branch-router rollouts, help desk for 500 branches.

Reference call under NDA

DEPLOYMENT HIGHLIGHTS · ACROSS ANCHORS

216

Branch routers deployed for a single government rollout · multi-ISP load balancing

500

Branches served by our help-desk automation across regulated infrastructure

3 ph.

AutoProvision phases at the tier-1 Thai telco · 2019 · 2020 · 2026

11+

Years of consecutive infrastructure engagement with our shortest anchor relationship

CASE STUDIES

We don't publish case studies publicly — most customers run regulated infrastructure where vendor disclosure is sensitive. Reference calls available during sales process, under NDA.

Request reference call →

CAREERS