Data Plane EngineerStealth Mode Startup
Baner, Pune · hybrid · 0-12 yrs · 1-45 LPA
Our core team is building proprietary generative AI infrastructure focused on efficient model performance. You will be instrumental in the end-to-end lifecycle of SLMs, moving from architectural design to deployment-ready weights. This role requires a hands-on approach to refining model outputs for high-precision, domain-specific tasks.
what you'll do:
- Design and implement training pipelines for small language models using open-weight architectures.
- Perform fine-tuning operations to optimize models for specific accuracy and performance benchmarks.
- Develop rigorous evaluation frameworks to measure output quality, relevance, and hallucination rates.
- Conduct deep-dive analysis on transformer attention mechanisms to improve model efficiency.
- Integrate techniques like quantization and pruning to enable deployment on constrained compute environments.
- Collaborate with internal teams to establish data pipelines that curate high-quality training datasets.
why join: You will have the opportunity to build foundational AI systems from scratch in a high-focus environment with complete ownership over the model training stack.
message me directly on tal BOSS.
tal saysnetworking data plane development engineer. handson with c, c++, go, linux, open source, tcp ip stack. understanding of
engineering
junior
hybrid
Baner, Pune
pythonpytorchjaxhugging-facecuda
comments 1 anon