Dmytro Dzhulgakov(@dzhulgakov):Training API is in GA: specialized intelligence with full flexibility algo: SFT, DPO, RFT, OPD, your own model: 1B to 3T efficiency: sync or async numerics: aligned training/rollout/inference pricing: start serverless (per-token, LoRA), ramp up to dedicated (scale, full-param)

Dmytro Dzhulgakov

@dzhulgakov

Co-founder and CTO @FireworksAI_HQ. PyTorch core maintainer. Previously FB Ads. Ex-Pro Competitive Programmer

参加 April 2013

829 フォロー中 7.6K ファン

Dmytro Dzhulgakov@dzhulgakov

2026.08.31 23:57

Training API is in GA: specialized intelligence with full flexibility algo: SFT, DPO, RFT, OPD, your own model: 1B to 3T efficiency: sync or async numerics: aligned training/rollout/inference pricing: start serverless (per-token, LoRA), ramp up to dedicated (scale, full-param)

Fireworks@FireworksAI_HQ

2026.08.31 15:33

The most ambitious companies are training models to outperform the frontier on the capabilities that differentiate their business. Today, we announce the general availability of our Training API and Fireworks Lab, making model specialization accessible to all organizations.