Seed Round 2025
Invest in the AI infrastructure
for engineering hiring.
EngineIQ is the first ML platform that scores real engineering work — PCB layouts, schematics, CAD geometry, and SPICE simulations — with theoretical confidence bounds. We are raising a $750K seed round.
Current traction
0.942Spearman r (ML accuracy)
49Production API routes
192Feature dimensions
32/32Integration tests pass
What's already built
Production ML scoring model
GBM ensemble, r=0.942 Spearman, 6 competency heads, 51ms CPU inference. Trained on 6,800 synthetic + fault-injected designs.
SRHN v4 reasoning engine
200KB C11 binary. PAC-Bayes theoretical bounds [R7], causal chains, contradiction detection, self-growing Hebbian graph. Port 8765 REST.
Real KiCad parser + DRC
Full S-expression parser for .kicad_pcb and .kicad_sch. 17 IPC-2221 rule violations, component-level localization, specific fix strings.
Multi-modal fusion (192-dim)
PCB (48) + schematic (32) + SPICE simulation (24) + MCAD geometry (48) + interaction trace (32) + SRHN reasoning (8). All wired end-to-end.
49-route production API + 20-page React SPA
FastAPI, Docker Compose, CI/CD via GitHub Actions, Prometheus metrics, JWT auth. CORS. 4-worker uvicorn config. Zero mock endpoints.
GATv2 GNN + Circuit LM (code complete)
6-layer GATv2, 256-dim, 3-objective pretraining. 85M Circuit Transformer with RoPE. Both need GPU time — first use of seed funding.
$750K
Seed round — 12-month runway
- 33%GPU compute (GNN + LM training)
- 27%Expert annotators (3 senior EEs)
- 24%Go-to-market + enterprise pilots
- 10%Security hardening (SOC2 Type II)
- 6%Legal, ops, infrastructure
For engineers
Collaborate and earn equity
We need senior engineers to validate and improve our scoring model. Meaningful contribution = equity stake + recognition in the platform.
Senior PCB Engineers
Annotate 70+ real PCB designs. Score: signal integrity, power design, layout quality, thermal management. Provide written rationale per design.
Expert-level KiCad / Altium experience
Familiar with IPC-2221, IPC-2141
5+ years production PCB design
Apply →
Analog / Power Engineers
Review power supply, motor controller, and RF designs. Evaluate component selection, protection circuits, and layout for EMI compliance.
Buck/boost/flyback design experience
EMC / IEC-61000 knowledge
SPICE simulation proficiency
Apply →
ML / AI Researchers
Contribute to GATv2 pretraining, Circuit LM architecture, or SRHN research modules. Open-source contributions welcome on GitHub.
Graph neural network experience
Transformer / LLM pretraining experience
Python + PyTorch proficiency
GitHub →
Investor FAQ
Common questions
What is the competitive moat?+
Three layers: (1) EE-native 192-dim feature vocabulary — not text embeddings, not general-purpose vision; (2) SRHN v4 reasoning engine with PAC-Bayes bounds — no competitor has theoretical confidence guarantees; (3) Behavior bonus from interaction trace — scores process quality, not just artifact quality, making scores very hard to game.
What is the business model?+
B2B SaaS: per-assessment pricing for enterprises ($15–45/assessment), platform licensing for bootcamps and certification providers, and API access for ATS integration. Year 1 target: 3 enterprise pilots at $5K–15K/yr.
What files does it currently support?+
.kicad_pcb and .kicad_sch (KiCad 6+), SPICE netlists (.net, .cir), and mechanical CAD (.step, .stl, .obj). Altium .PcbDoc and Eagle .brd parsing is on the roadmap.
How does the GNN training fit in?+
The GATv2 GNN is the primary use of GPU funding. Once pretrained on 84k+ real KiCad schematics, the 256-dim learned graph embedding replaces hand-crafted PCB+schematic features. This makes scores dramatically harder to reverse-engineer and catches circuit-topology patterns no feature engineering can express. Estimated r → 0.97+ after GNN fine-tuning.
Is the model already working?+
Yes — the GBM model is live and achieves Spearman r=0.942 on validation data. 32/32 integration tests pass. The Docker Compose stack deploys in one command. The GNN and Circuit LM are code-complete but untrained (need GPU). See the pitch deck for a live demo walkthrough.
How do I get in touch?+
Email srhn.edge@outlook.com with "Investment Enquiry" or "Collaborator Application" in the subject line. We respond within 24 hours. For technical due diligence, request GitHub access to the private repo.