Deskripsi Pekerjaan
Nexus Intelligence Labs is at the forefront of applied AI research, building foundation models that redefine enterprise automation. We are seeking a Senior Machine Learning Engineer to lead the lifecycle of large-scale generative models—from architecture design and distributed training to low-latency production deployment.
You will join a tight-knit team of researchers and engineers who value intellectual honesty, rapid iteration, and open-source contribution. This role offers significant autonomy to shape our technical roadmap and the compute infrastructure backing our next-generation models.
Why Join Us?
- Frontier Compute: Direct access to large GPU clusters (H100s/A100s) via strategic cloud partnerships.
- Research-to-Production Velocity: Push code to production impacting millions of users within days, not quarters.
- Equity Upside: Pre-IPO stock options with a transparent valuation structure.
- Learning Budget: $5,000/year for conferences, courses, and certifications.
Tanggung Jawab
- Design, implement, and optimize distributed training pipelines for LLMs and multimodal models using PyTorch, FSDP/DeepSpeed, and Megatron-LM.
- Architect scalable MLOps infrastructure (feature stores, model registries, CI/CD for ML) on Kubernetes (EKS/GKE).
- Lead model quantization, compilation (TensorRT, Triton), and serving optimization for sub-100ms latency at high throughput.
- Collaborate with Research Scientists to translate novel architectures (Mixture of Experts, RAG, Agentic workflows) into robust production systems.
- Establish rigorous evaluation frameworks: offline benchmarks, A/B testing infrastructure, and human-in-the-loop feedback loops.
- Mentor junior engineers, conduct design reviews, and champion best practices in code quality, reproducibility, and GPU utilization.
- Contribute upstream to open-source ecosystems (PyTorch, Hugging Face, vLLM) and publish findings at top-tier venues (NeurIPS, ICML, ICLR).
Kualifikasi
- M.S. or Ph.D. in Computer Science, Machine Learning, or equivalent practical experience (5+ years industry ML engineering).
- Expert-level Python and deep mastery of PyTorch internals (autograd, custom kernels, distributed comms).
- Proven track record training and deploying models >10B parameters on multi-node GPU clusters.
- Strong systems engineering background: Linux kernel tuning, CUDA programming, networking (InfiniBand/RoCE), and container orchestration.
- Experience with LLM-specific tooling: Hugging Face Ecosystem, LangChain/LlamaIndex, Vector DBs (Milvus, Pinecone, Weaviate).
- Demonstrated ability to debug non-deterministic training runs and optimize memory/throughput trade-offs.
- Excellent communication skills; ability to distill complex technical concepts for cross-functional stakeholders.