# Kostiantyn Polosukhin

Kostiantyn Polosukhin is a software-company CEO focused on manufacturing automation and industrial AI. Their work also covers post-quantum cybersecurity, including HarvestGuard, a tool for finding cryptography that may need migration.

- Profile: https://b2binfluencer.marketing/creators/kostiantyn-polosukhin/
- LinkedIn: https://www.linkedin.com/in/kostiantyn-polosukhin-957b6a26
- LinkedIn followers: 3,889
- Topics: artificial-intelligence, industrial-ai, industrial-automation, post-quantum-cryptography, ai-agent-security
- Based on 18 dated posts, latest 2026-10-07

## Summary

Kostiantyn Polosukhin is most useful for explaining manufacturing automation, industrial AI, and post-quantum migration as operational problems.

- Best for: Manufacturing technology brands selling integration, industrial data platforms, industrial AI, or post-quantum readiness to factories.
- Not for: Consumer travel or mass-market financial services brands seeking high-reach awareness on booking fraud and phishing.
- Check first: Confirm recent analytics and verify company service claims before contacting them.

### Takeaways

- Quantum risk is a present-tense data problem, not a future one.
- The cryptography that matters is often code nobody wrote.
- Factories need trustworthy data before they need AI.

## Main ideas

- **Quantum risk is a present-tense data problem, not a future one**: Polosukhin’s repeated argument is that encrypted data captured now is being stored for decryption once a capable quantum computer exists — “harvest now, decrypt later.” On that basis, migration is a current data-handling problem, not a future research concern. Exposure lies in source code, configuration files, dependencies, certificates and framework defaults, not just network traffic. They frame the deadlines as supply-chain obligations: under the US executive order they cite, key establishment is due by 31 December 2030 and authentication by 31 December 2031, with federal contractors covered alongside agencies. Organisations that start early retain architectural options; those that wait risk turning architecture decisions into compliance deadlines.
- **The cryptography that matters is often code nobody wrote**: Polosukhin argues that the hardest part of post-quantum migration is architecture visibility rather than cryptography itself, because standard frameworks introduce quantum-vulnerable cryptography silently. Findings from their own scan are grouped into three categories — expected, forgotten and invisible — where "expected" algorithms such as RSA and ECDSA are at least known, "forgotten" configuration files and certificates from untouched dependencies sit in production unnoticed, and "invisible" cryptography is negotiated by frameworks without a single line of cryptographic code. They contrast this with tools that never open source files, tools that report findings without saying what to replace them with, and AI-assisted tools that send code to a cloud API, creating a new exposure while removing one.
- **AI changes cryptographic design, and the answer is crypto-agility rather than retreat**: Reporting on an AI-assisted break in a post-quantum candidate, Polosukhin argues that two races are now underway: quantum computers against encryption, and AI against cryptographic design. AI is well suited to the exhaustive structural search that catches lattice weaknesses. They argue that ML-KEM, ML-DSA and SLH-DSA are not the withdrawn candidate, and call for crypto-agility: systems that let teams swap algorithms without rewriting the architecture.
- **Factories need trustworthy data before they need AI**: Polosukhin’s core manufacturing argument is that a self-running factory depends first on accurate data and systems that can communicate. In their account, the obstacle is rarely the machines: manual entry creates different formats, naming conventions and undetected typos, making production data less reliable than people assume. That is why many AI initiatives struggle long before the model becomes the problem. Without trusted, connected data available in real time, there is nothing reliable to learn from or act on. They argue that most manufacturers are closer than they think, and that competitive advantage lies in the gap between interest in AI and operational readiness.
- **The pattern in industrial technology is integration, not hype**: Polosukhin runs a stated weekly format of five signals across industrial automation, AI, robotics, additive manufacturing and deep tech. Each item explains what changes on the factory floor rather than treating the development as an announcement. Their closing thesis: winners will connect new technology to existing infrastructure, production data, safety requirements, operators, maintenance teams and long-term reliability.
- **Agent capability needs a security layer built at the same time**: Starting with an incident in which an AI agent worked around access restrictions on a government portal, Polosukhin argues that the security question shifts once agents can act: from what a model can say to what it is allowed to do. They propose an on-premises gateway for autonomous AI agents that validates tool calls before execution, stops behaviour outside defined boundaries, and masks personal and health data inside the client’s environment. Their case is for building the security layer alongside the intelligence, not against agents.
- **Combining and deploying models is the scarce skill**: Polosukhin reads a wave of senior departures from OpenAI and Google as the end of the platform phase and the start of vertical solutions. They argue that the scarce skill is not building foundation models but combining them and putting them into production systems in specific industries. In their account, value comes from engineers who understand the business problem beneath the technology and work with agentic workflows, multimodal reasoning, real-time data integration and domain-specific fine-tuning.
- **Credibility comes from naming partners and communities in public**: Across two posts, Polosukhin credits other organisations by name and presents open knowledge-sharing as a reason to engage with a community. On partnerships, they say plainly that when a client or partner needs expertise their company lacks, they would recommend the partner’s team without hesitation and publish the joint work. On community, they single out the decisions that make an ecosystem useful — free training, shared templates and direct access to the people who built the platform — rather than treating a conference as a sales channel.

## Work to read

- [5 manufacturing tech signals worth watching](https://www.linkedin.com/posts/kostiantyn-polosukhin-957b6a26_5-manufacturing-tech-signals-worth-watching-activity-7480276485866602497-DZd_)
- [Speed-Running the Next Industrial Revolution: AI, Additive Manufacturing, and the Factories That Build Themselves](https://www.linkedin.com/posts/kostiantyn-polosukhin-957b6a26_manufacturing-industrialautomation-londontechweek-activity-7473381682927169536-iN6n)
- [ICC is back in Sacramento this September](https://www.linkedin.com/posts/kostiantyn-polosukhin-957b6a26_icc2026-industrialautomation-ignition-activity-7488607738101604352-Qyx9)
- [From gene therapy manufacturing to industrial AI — four years of partnership with InflexionPoint](https://www.linkedin.com/posts/kostiantyn-polosukhin-957b6a26_industrialautomation-industrialai-partnership-activity-7511440263626637313-XIdm)
- [Harvest now, decrypt later: travel cyberattacks and quantum risk](https://www.linkedin.com/posts/kostiantyn-polosukhin-957b6a26_postquantum-cybersecurity-harvestguard-activity-7490774953491894272-sYji)
- [An AI model broke a post-quantum cryptography candidate in 60 hours](https://www.linkedin.com/posts/kostiantyn-polosukhin-957b6a26_postquantum-cybersecurity-harvestguard-activity-7490418951848353792-oOvp)

## Brand relationships

- **HebronSoft** (employment)
- **HarvestGuard** (owned_business)
- **InflexionPoint** (disclosed_partnership)
- **Inductive Automation** (mention)

## Partnerships

Kostiantyn Polosukhin does not publish sponsorship packages, rates, or a media kit. The posts describe delivery partnerships: HebronSoft works alongside other engineering firms, names them publicly, and publishes case studies on joint projects. Brand-side relationships are client engagements rather than paid placements, and every new engagement includes a HarvestGuard scan. Brands should expect to scope a project directly with the team, not book a self-serve advertising slot. For partnership inquiries, contact Polosukhin through LinkedIn or HebronSoft through its website; ask for pricing, deliverables, and usage rights.

## Alternatives

- Nick Sullivan: Choose Nick Sullivan when the campaign targets security engineers, cloud platform owners, or web-infrastructure decision-makers asking how to deploy post-quantum cryptography at internet scale. His posts describe Google switching on post-quantum cryptography for internal communications and then in the Chrome browser, and Cloudflare shipping compatible cryptography for internal traffic and origin-server connections. He argues that by the end of that year, a substantial share of web traffic could be protected this way. He casts this as the transition's early-adopter phase, points businesses to NIST guidance for planning, and offers consultations to organizations working through it. He also covers broader security topics: confidential computing and software attestation, how privacy proxies such as iCloud Private Relay are weakening IP-reputation-based detection, and the growing complexity and cost of production security scanning. That makes him a stronger pick than Polosukhin for campaigns about protocols, browsers, and cloud-scale migration, or security-leadership advice rather than plant-floor issues.
- Bryant D Nielson: Choose Bryant D Nielson when the campaign targets board, risk, audit, and governance buyers rather than engineering teams. His posts center on a named governance framework, QCI-QS1 v2.3, and its Q-Risk Score, including a data longevity and exposure pillar that turns encryption risk into a retention-policy question: if data lifetime plus migration time exceeds the time remaining before cryptographic disruption, the organization already has a problem. He argues that the real deadline often lies in policies requiring confidentiality for 7, 10, 15, or 20 years, and asks which information held today will still need protection when its cryptography is no longer sufficient. He is also convening a State of Quantum Readiness Summit in Arlington, Virginia, on 10 March 2027, bringing cybersecurity, audit, standards, risk, and governance people together without vendor keynotes. That makes him a better fit than Polosukhin for campaigns about how to inventory, measure, evidence, and govern quantum risk, or for reaching non-technical decision-makers and auditors.
- Alex Pruden: Choose Alex Pruden when the campaign targets digital-asset holders, exchanges, custodians, or crypto-adjacent financial buyers rather than factories. He approaches quantum risk from the asset side. He compares IonQ's estimate of what it would take to break Bitcoin's elliptic-curve signatures with earlier estimates from Google and Oratomic, arguing that independent teams converging on compatible numbers is itself a signal. He explains the tradeoffs among superconducting, trapped-ion, and neutral-atom machines: Google's design needs under 500,000 physical qubits but finishes in minutes, while IonQ's needs roughly 20,000 but takes about 26 days per attempt. He draws an operational distinction from that difference: an attack measured in minutes could target a transaction while it is still being confirmed, while a 26-day attack threatens long-lived exposed keys and static balances such as Satoshi's early coins. He also writes a plain-language primer on neutral-atom hardware, from trapping atoms in laser beams to the first error correction on the hardware and a 2028 fault-tolerant roadmap. He says Project Eleven is rebuilding digital-asset custody to survive a post-quantum world. For campaigns about why already-exposed keys need migrating now, or explaining quantum hardware to non-specialists, he is a better choice than Polosukhin.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/