Centella AI Therapeutics is an AI-native drug discovery company founded by practicing scientists with 30+ years of combined experience. Our platform, Centella CASCADE™, is an end-to-end, disease-agnostic engine that blends generative models, medicinal chemistry to design high-quality, drug-like molecules rapidly and intelligently. We are seeking curious and technically strong fellows who are passionate about building AI that engineers molecules, not just models
Key Responsibilities:
- Architect and build end-to-end Agentic AI systems across cloud and on-prem. •
- Integrate LLMs, cognitive services, and ML frameworks like Semantic Kernel, LangChain, and AutoGen.
- Build Retrieval-Augmented Generation (RAG) pipelines using vector databases and knowledge stores.
- Enable agents to interpret and act on chemical and biomedical datasets.
- Deploy scalable agents via Kubernetes, serverless, or containerized services.
- Optimize LLMs through prompt engineering, fine-tuning, and caching.
- Monitor system performance using observability and logging tools.
- Iterate agent workflows based on testing, analytics, and scientific insights.
- Collaborate with AI/ML teams and scientists to align with research goals.
- Iterate workflows and maintain documentation following Responsible AI practices
Qualifications:
- 1–2 years of experience in Machine Learning, Deep Learning, or computational drug discovery
- Bachelor’s/Master’s in CS, AI/ML, Bioinformatics, Comp. Biology, ChemE, or related field
- Strong Python + experience with PyTorch/TensorFlow model development
- Exposure to generative models (Diffusion/VAEs/Transformers) - academic or industry
- Experience designing and deploying AI workflows on multi-cloud or hybrid infrastructure
- Familiarity with vector databases, RAG pipelines, and autonomous agent-driven workflows
- Knowledge of multi-agent orchestration, memory management, security, and Responsible AI principles
- Understanding of security and compliance in multi-cloud environments
Logistics:
- Duration: 3 to 6 months
- Type: Onsite
- Location: Hyderabad
