Senior Software Engineer

You own the hard latency engineering. Sub-100ms is not a target for you — it's a prior. You

have lived inside systems where the latency budget is a commercial constraint, not an

engineering aspiration. You know what breaks at 10M events per day because you've watched

it break.

You report to the Founding CTO. You bring the production execution depth that the protocol

requires in its first deployment. The Founding CTO designs the intelligence that produces the

score. You make that intelligence fast, scalable, and cryptographically anchored.

This is a senior engineering role with meaningful equity and significant influence on architectural

decisions within your domain. You are joining a founding team at the earliest stage.

WHAT YOU OWN

L0 — Connector Layer

How AUTHX plugs into partner stacks. OpenRTB 2.x bid stream integration. Endpoint

placement between DSP and SSP. Webhook-based call sequence. Sub-100ms response

requirement end-to-end. SDK connectors. Authentication, rate limits, error handling, retry

logic under SSP auction constraints.

Real-Time Architecture

Redis cache for score lookup under 10ms. Hash verification under 5ms. Tier check under

1ms. Edge deployment specification. The sub-100ms hot path is your primary engineering

responsibility. You architect for RTB-style bid volume spikes during prime-time.

L4 — Attestation & Fee Pipeline

Batch size, batching interval, L2 chain candidates (Polygon, Arbitrum, or equivalent based

on gas economics at deployment), gas optimisation, Merkle proof retrieval for audits. The L4

pipeline that makes the protocol's economics real.

Critical architectural constraint: Fee is computed from tier, downstream of scoring. It is

never an input to the score. The acyclic boundary (Score → Tier → Fee) is architectural and

you enforce it. L4 proves the score — it does not contribute to the score.

L2 Inference Hardening

While the Founding CTO owns model selection and Fracture Index design, you own making

that inference run under 35ms in production. ONNX runtime, edge caching of embeddings,

quantization, hardware selection. The CTO designs the intelligence. You make it fast

enough to ship.

Production Hardening

Taking the proven architecture from specification and reference implementation to 10M+

QualificationEvents per day with zero-fault data integrity. RPO = 0. RTO ≤ 30 minutes.

99.99% uptime per partner per month.

WHAT YOU ARE NOT RESPONSIBLE FOR

The AI layer. Entity Identity extraction, Fracture Index computation, semantic inference model

design — these belong to the Founding CTO. You interface with the outputs of the AI layer

(scores, tier assignments, embeddings) but you don't build the intelligence that produces them.

DAB (Digital Autonomous Builder). The entity onboarding engine is the CTO's domain. You

consume Entity Identity from the registry. You don't build the extraction pipeline that creates it.

BACKGROUND

10+ years in production infrastructure at companies that earn on transaction flow.

Most likely backgrounds:

• RTB / Adtech infrastructure — IAS, DoubleVerify, The Trade Desk, AppNexus/Xandr,

Magnite

• Payments infrastructure — Stripe, Adyen, Braintree, Plaid engineering teams

• Distributed ledger infrastructure — L2 chains, exchange backends, custody systems

• HFT-adjacent systems — exchange matching engines, market data infrastructure

• Protocol-layer API companies — Twilio, SendGrid, Bandwidth infrastructure teams

The rarity isn't any single skill. It's the combination of RTB-grade latency discipline AND

cryptographic/L2 engineering in one person. That combination is what makes this hire

critical-path.

You have built systems under hard SLAs where failure has commercial consequences. You

have designed for scale before it arrived. You understand that latency is a product requirement,

not an engineering preference.

WHAT SUCCESS LOOKS LIKE

• L0-L4 hot path running under 100ms P95 in production

• 10M+ QualificationEvents/day with zero data loss

• Merkle batching live on L2 chain with gas < $0.0001 per event

• First partner sandbox operational within 90 days of start

• Real-time architecture documented and reproducible

• 99.99% uptime per partner, per month

What we offer

• Competitive compensation aligned with principal infrastructure and distributed systems talent in your region.
• Direct collaboration with the Founding CTO, with full ownership of the architecture and technical direction.
• Remote-first working environment, with flexibility for candidates who can maintain good overlap with the London timezone.
• The opportunity to build core infrastructure that processes 10+ million events per day while maintaining a sub-100ms P95 latency budget.

To ensure information security and fully comply with the Law on Personal Data Protection No. 91/2025/QH15, Bestarion is committed to protecting your candidate data. By submitting your application on this platform, you acknowledge our Privacy Policy: https://bestarion.com/privacy-policy/

If your profile matches our requirements and you proceed to the interview stage, we will send an official email requesting your formal consent before further processing your data.