Raising a £1.2M seed to expand our benchmarking platform. £400k committed from angels.
Most teams overspend on GPUs because they skip workload profiling. We published our sizing methodology.
Pre-seed update: 2 paid pilots, 14-week sales cycles, path to £20k MRR by year end.
Open-sourcing a library for tracking GPU cost per request across multi-tenant serving.
Closing a small seed extension. Use of funds: first sales hire and SOC 2.
How are you handling prompt versioning in production? Curious what others use before building internal tooling.
Build log: cut p95 latency on our RAG pipeline from 4.2s to 1.1s with embedding cache per tenant.
Shipped v0.3 of our eval harness: regression tests on every PR with a frozen golden set.
Week 12 update: migrated to spot GPU instances with checkpointing. Cost down 38%.
Raising a £1.2M seed to expand our benchmarking platform. £400k committed from angels.
Pre-seed update: 2 paid pilots, 14-week sales cycles, path to £20k MRR by year end.
Closing a small seed extension. Use of funds: first sales hire and SOC 2.
Build log: cut p95 latency on our RAG pipeline from 4.2s to 1.1s with embedding cache per tenant.
Shipped v0.3 of our eval harness: regression tests on every PR with a frozen golden set.
Week 12 update: migrated to spot GPU instances with checkpointing. Cost down 38%.
Most teams overspend on GPUs because they skip workload profiling. We published our sizing methodology.
Open-sourcing a library for tracking GPU cost per request across multi-tenant serving.
How are you handling prompt versioning in production? Curious what others use before building internal tooling.
Raising a £1.2M seed to expand our benchmarking platform. £400k committed from angels.
Most teams overspend on GPUs because they skip workload profiling. We published our sizing methodology.
Pre-seed update: 2 paid pilots, 14-week sales cycles, path to £20k MRR by year end.
Open-sourcing a library for tracking GPU cost per request across multi-tenant serving.
Closing a small seed extension. Use of funds: first sales hire and SOC 2.
How are you handling prompt versioning in production? Curious what others use before building internal tooling.
Build log: cut p95 latency on our RAG pipeline from 4.2s to 1.1s with embedding cache per tenant.
Shipped v0.3 of our eval harness: regression tests on every PR with a frozen golden set.
Week 12 update: migrated to spot GPU instances with checkpointing. Cost down 38%.
Most teams overspend on GPUs because they skip workload profiling. We published our sizing methodology.
Open-sourcing a library for tracking GPU cost per request across multi-tenant serving.
How are you handling prompt versioning in production? Curious what others use before building internal tooling.
Raising a £1.2M seed to expand our benchmarking platform. £400k committed from angels.
Pre-seed update: 2 paid pilots, 14-week sales cycles, path to £20k MRR by year end.
Closing a small seed extension. Use of funds: first sales hire and SOC 2.
Build log: cut p95 latency on our RAG pipeline from 4.2s to 1.1s with embedding cache per tenant.
Shipped v0.3 of our eval harness: regression tests on every PR with a frozen golden set.
Week 12 update: migrated to spot GPU instances with checkpointing. Cost down 38%.