←
BeChained1,281 posts
BeChained avatar

BeChained

@be_chained

Our #AI tech to automatize #energyefficiency in #manufacturing #processes Unlock #energyflexibility, #CO2credits, #offgrid #capacity #Industry50

◷ Joined May 2019▧ 178 Photos and videos
Alex Iskold | 2048.vc avatar
⟳ BeChained retweeted
Alex Iskold | 2048.vc@alexiskold·Sep 3

What to do when a pre-seed investor is writing a sizable check and insists on a board seat ? First let’s discuss round and check size. If the round is $4m - $5m and check is $2m - $3m I can see this coming up. When the round is $1m and check is $250k it makes no sense. Regardless if the founders don’t have a choice and need to take the capital but don’t want to give a board seat but investors insist on it what to do ? The best compromise is give the seat but make it expire at the following financing. This should work well and is fair if investors insist on having an oversight until the next round. If the small check investor doesn’t agree and wants to hang on the board permanently this is a massive red flag 🚩 imo and the founders should consider walking away. If the investors agree to expiration at the next round it’s fine for the founders as they will continue to maintain control and will expand the board as the company scales.

BeChained avatar
BeChained@be_chained·Sep 10

Stefano Melchior is heading to hashtag#Montreal for ALL IN #2026 (allinevent.ai)! 🇨🇦✨ Thanks to @ACCIÓ for organizing this mission and to connectwith world-class AI ecosystems, including @scaleai , @Luqia, @IVADO_Qc , #Videns by @COFOMO, @VectorInst, @AmiiThinks

BeChained avatar
BeChained@be_chained·Aug 3

Can #RAG optimize #machinery without blowing up context? The data says yes—if for #industry. - Text-only baseline: At 128K context, the baseline's accuracy dropped to less than 28% - The Multimodal Pipeline: at 128K context split,#IndyRAG achieved accuracy of 85.4% (semantic LLM)

BeChained avatar
BeChained@be_chained·Jul 23

🔹 Checkpoint Recovery: Resumes execution on network/GCP drops—no restarting from scratch. 🔹 Self-Healing Loop: 3-tier automated error handling that auto-adjusts queries to fix blockers.

BeChained avatar
BeChained@be_chained·Jul 23

We’re making industrial Digital Twin generation faster, automatized, and fully resilient: 🔹 Layered RAG: Domain-focused agents for Machine Ops, Energy, and Physical Layout, 🔹 Dynamic Glossaries: RAG adapts automatically to process-specific tech jargon,

BeChained avatar
BeChained@be_chained·Jul 9

2) Capable of parsing dense, multi-language ISO A0 drawings & technical manuals 3) Built generalized skills/hooks to map raw #PLC/#SCADA signals (with #OPCUA) from legacy hardware 4) Standardized @gitlab CI/CD hybrid deployment to keep sensitive OT data strictly on-prem

BeChained avatar
BeChained@be_chained·Jun 8

Key focus areas for this deployment: 🔹 Tech Validation: Scaling in a demanding automotive supply chain environment. 🔹 Automated Rollout: Streamlining how we clone industrial lines even faster. 🔹 TRL 6 to TRL 7: Moving from prototype to full-scale operational demonstration.