AI Security Research Tooling

Enterprise LLM Infrastructure
for Security Teams

Apex Solutions builds open-source LLM inference tooling and AI copilots that power security researchers and enterprise red teams — enabling automated vulnerability analysis, penetration testing assistance, and security audit automation at scale.

2 Live Products
19+ Security Tools
164B Param Support
75–99% AWS Workloads
Built on Python llama.cpp / ds4 SkyPilot AWS & GCP GGUF / CUDA OpenAI-compatible API A100-80GB / H100 DeepSeek V4 Flash Qwen3 35B
Products

AI Tooling Built for Security

Our products combine state-of-the-art open-source LLMs with purpose-built tooling designed around the security research workflow.

☁️

Cloud Inference Platform

Live · Expanding

Managed GPU orchestration for running large GGUF models (up to 164GB) across AWS, GCP, and Azure. Uses SkyPilot for multi-cloud scheduling with preemptible instance optimization — targeting A100-80GB and H100 at $1.52–$2.81/hr spot pricing.

Multi-Cloud SkyPilot Spot Optimized A100 / H100 OpenAI API
Platform

From Model to Security Insight

A vertically integrated stack — from GPU infrastructure to domain-specific security tooling.

Step 01
🖥️

Model Deployment

Quantized GGUF models deployed on GPU instances via SkyPilot. Runtime VRAM detection sets optimal layer offloading per hardware class (A100-80GB → 40 GPU layers, fallback → 20).

Step 02
🔌

OpenAI-Compatible API

ds4-server exposes a standard REST endpoint on port 8000. Drop-in replacement for any client using the chat completions API. LiteLLM proxy handles multi-model routing.

Step 03
🛠️

Security Tool Integration

19 security tools wired to the agent runtime — recon, exploit analysis, CVE lookup, payload generation, binary analysis, and structured report generation.

Step 04
📋

Structured Audit Output

CVSS-scored findings, remediation guidance, and audit reports delivered via terminal, web browser, mobile app, desktop app, or API. Fully reproducible — no cloud dependency.

Use Cases

Built for Enterprise Security Workflows

Our customers are security researchers and enterprise teams that need LLM-powered assistance without compromising sensitive client data.

🎯

Red Team Engagements

AI-assisted attack path analysis, payload generation, and lateral movement planning for enterprise red team operations.

🔍

Vulnerability Analysis

Automated triage and CVSS scoring of vulnerability findings. Correlates CVE databases with target system context.

📄

Audit Report Generation

Structured, client-ready security audit reports generated from raw findings with remediation recommendations.

🔒

Air-Gapped Research

Run frontier-class LLMs on local GPU hardware. No data leaves the environment — suitable for classified or regulated engagements.

🤖

Security Automation

Automate repetitive pentesting tasks — reconnaissance, enumeration, and initial analysis — with AI-driven terminal agents.

☁️

Cloud-Scale Inference

Deploy large models on AWS/GCP spot instances for high-throughput batch analysis at a fraction of on-demand pricing.

About

Who We Are

The Problem We Solve

Most AI security tools rely on closed, cloud-hosted LLMs that cannot process sensitive client data. Security researchers working on enterprise engagements need the intelligence of frontier models — but without the data sovereignty risk that comes from sending client environments to a third-party API.

Our Approach

Apex Solutions was founded in January 2026 to close that gap. We build on open-source models — DeepSeek V4 Flash and Qwen3-35B-A3B — deployed on customer-controlled GPU infrastructure. Security teams get frontier-class intelligence without the compliance risk.

Our cloud inference platform uses SkyPilot to orchestrate GPU workloads across AWS, GCP, and Azure — automatically selecting preemptible instances to minimize cost. The same GGUF models run identically on local RTX hardware or a cloud A100-80GB, making development and production environments truly portable.

Where We Are Now

We are bootstrapped, generating revenue from security research engagements, and iterating toward a broader platform. Our products are in active use for enterprise red-team work, with 75–99% of our infrastructure spend on AWS. We are a 1–10 person team based in Kolkata, West Bengal, India, operating since January 2026.

Python llama.cpp / ds4 SkyPilot AWS EC2 / P5 GCP A100 GGUF / CUDA OpenAI API LiteLLM HuggingFace Hub DeepSeek V4 Flash Qwen3-35B
S
Soumabha Majumdar
Founder & CEO

Security engineer and AI researcher building open-source LLM infrastructure for enterprise security teams. Founded Apex Solutions in January 2026 to bring frontier AI capabilities to security research workflows without compromising data sovereignty.

📍 Kolkata, West Bengal, India
✉️ admin@apexml.xyz
🏷️ CEO · Entrepreneur / Co-Founder
Jan 2026 Founded
Bootstrapped Funding Stage
1–10 Team Size
AI / SaaS Industry
Contact

Get in Touch

Interested in our products, a security research engagement, or partnership? Reach out directly — we respond within 24 hours.

✉️
Email admin@apexml.xyz
🌐
Website https://apexml.xyz
📍
Location Kolkata, West Bengal 700030 · India
🏢
Company Apex Solutions — AI Security Research Platform