Free External Exposure Snapshot — no obligation

Find your company's external exposure
before an attacker does

7+ years in enterprise cybersecurity — SIEM optimization, incident response, threat intelligence — now applied to AI-era security. Request a free passive snapshot of your organization's publicly visible exposure, no strings attached.

Request Free Snapshot Services Projects
Saer Ndiaye
7+Years in Cybersecurity
4Industry Certifications
10+AI Automation Projects
Security-FirstEngineering Approach

Cybersecurity • AI Security • Automation Engineering

Services

Security-first engineering for your business and your AI systems

Grounded in enterprise defensive security and authorized security testing — not just prompt-engineering best practices.

Free External Exposure Snapshot

A no-cost, no-obligation snapshot of what's publicly visible about your company's attack surface — domains, DNS, exposed certificates, email-security posture (SPF/DKIM/DMARC), and related signals. Fully passive: no scanning your live systems, no risk to your infrastructure. You get a handful of high-level findings, with an invitation to go deeper in a full assessment if it's useful.

Authorized AI Security Assessments

Controlled testing of LLM-integrated applications to identify security weaknesses before they reach production.

AI Risk Assessments

Structured evaluation of AI system exposure — data handling, model access, and failure blast radius.

Secure LLM Integrations

Architecture review and hardening for systems wiring LLMs into production data and business logic.

AI Security Consulting

Advisory for teams shipping AI features who need a security-first second opinion before launch.

Passive Exposure Research

Publicly-sourced domain, DNS, certificate, and email-security signal analysis — see Africa Recon.

SOC Automation

SOAR-style playbooks and AI-assisted triage to reduce analyst response time without losing oversight.

AI & Workflow Automation

Design and automation of business workflows using AI, APIs, and system integrations — reducing repetitive work while maintaining monitoring, reliability, and human oversight. Examples include: intelligent alerts, executive briefings, lead qualification, customer-support workflows, and operational automation.

Featured Work

Projects

Selected systems and prototypes spanning production automation, internal tooling, and security research. Each one is built with explicit failure handling, not just to work in a demo.

About

Cybersecurity Foundation. AI-Driven Future.

I spent my career defending enterprise networks before I started building the systems I now automate with AI. That order matters — I design workflows with explicit failure handling, monitoring, and safe fallback behavior, not just workflows that demo well.

Senior Security Analyst experience covers SIEM optimization, incident response, and threat intelligence at enterprise scale. That background now shapes how I build AI systems: monitored, auditable, and designed to degrade gracefully instead of failing silently.

  • Built an enterprise Splunk investigation dashboard that reduced daily investigation effort by several analyst hours
  • Building an independent CSOC investigation-assistant prototype to explore AI-assisted analyst triage
  • IBM Generative AI certified, with additional AI security credentials
  • Extensive experience improving and automating security and business workflows

Today I split my time between hands-on security analysis and building AI systems — voice agents, workflow automation, SOAR-style playbooks, and structured pipelines that connect LLMs to real business processes.

  • Voice AI agents for real-world conversational workflows
  • Workflow automation connecting APIs, webhooks, and internal tools
  • Structured JSON pipelines for reliable LLM-to-system integration
  • Production monitoring, observability, and alerting — not just happy-path demos
How I Build

Engineering Principles

The difference between an AI demo and an AI system that survives contact with production.

01

Security by design

Every system is built assuming it will be attacked or misused, not just used correctly.

02

AI with human oversight

LLMs propose, humans (or explicit guardrails) approve anything consequential.

03

Reliability over demos

A workflow that works 95% of the time silently is worse than one that fails loudly 5% of the time.

04

No silent failures

Production-facing automations should include monitoring, actionable alerts, and appropriate fallback behavior.

05

Evidence-driven decisions

Ship, measure, and let real usage data — not intuition — drive the next iteration.

06

Modular, maintainable systems

Components that can be replaced, upgraded, or debugged in isolation, not tangled monoliths.

Process

How I Build

  1. Understand the problem.Talk to the actual users of the system before writing a line of code.
  2. Design the architecture.Map data flow, failure modes, and security boundaries up front.
  3. Build an MVP.The smallest version that proves the core value, not the full feature set.
  4. Test with real data.Synthetic tests don't surface the edge cases that matter.
  5. Add monitoring and failure handling.Monitoring and resilience before it ever touches production traffic.
  6. Iterate based on evidence.Let logs and outcomes — not assumptions — decide what changes next.
Technical Stack

Tools & Technologies

AI

  • OpenAI
  • Claude
  • Gemini
  • Retell AI / Vapi

Automation

  • n8n
  • Make
  • Zapier
  • APIs & Webhooks

Security

  • Splunk
  • Microsoft Defender
  • SOAR
  • Kali Linux

Infrastructure

  • Python
  • Linux
  • Docker
  • Git / Supabase
Contact

Request your free External Exposure Snapshot

Tell me a bit about your company and I'll send back findings — usually within a few business days.

LinkedIn