D
Deccan AI builds evaluation infrastructure for large language and multimodal models, serving frontier AI labs and enterprises. The stack—Python, PyTorch, Hugging Face, FastAPI, plus distributed systems (Kubernetes, Kafka, Ray)—reflects a company deep in training and inference optimization. Engineering dominates the hiring mix (7 of 10 active roles), concentrated at mid and senior level, with concurrent pain points around distributed systems depth and large-scale data pipeline optimization. This signals a team scaling infrastructure faster than specialized talent can be sourced.
Tech Stack 63 technologies
Core StackPython Java Go AWS Kubernetes PyTorch Hugging Face FastAPI Docker Snowflake Kafka RabbitMQ C++ TensorFlow HubSpot Zapier React JavaScript GCP Transformers BERT Google DeepMind Azure Ray JAX DeepSpeed vLLM TensorRT Triton Apollo+33 more
What Deccan AI Is Building
◆Challenges
- Optimizing distributed training performance
- Lack of deep distributed systems engineers
- Need engineers with depth and breadth
- Complex cloud infrastructure and security stack
- Scaling internal tooling
- Operational pain points
- Scaling core systems for ai training
- Optimizing large-scale data pipelines
- Improving system scalability and reliability
- Scaling rl training infrastructure
▲Active Projects
- Distributed ai systems powering rl training pipelines
- Cloud infrastructure on aws or gcp
- Infrastructure as code and ci/cd pipelines
- Multi-agent ai systems
- Rag pipelines
- Nlp chatbots
- Internal platforms and automation systems
- Ai evaluation pipelines
- Workflow orchestration
- Ai training and evaluation platforms
Hiring Activity
10 roles · 9 in 30d
Department
Seniority
Company intelligence