AI/ML/Gen AI Architect

Overall 10+ yrs of design and architecture experience, Min. 5+ years of experience in related area (AI/ML, Gen AI combined)in design and architecture of production (customer or internal) or customer lab deployed solution.
3. Good understanding of software design patterns, UML, application development (python, spark) best practices, SDLC exposure.
4. Understanding of Open API specification and REST APIs (Sync and Async APIs) in context of AI/ML applications.
5. Relational, Non-relational database modelling.
6. Understanding  and solution design experience of multi agent solutions, architectures (ReAct, Reflexion etc.), Tool calling, prompting techniques (chain of thought etc).
7. Experience with LLMOPs technology stack and applying them to real-world use cases (chat bots, text to SQL etc.)
  a. Integrating with Generative AI models (e.g., GPT-4, Mistral, Llama etc.)

     - understanding of Concepts of Quantization, KV caching, transformer architecture etc.
  - LLM deployment (using NIM, vLLM, Ollama etc.) and configuration experience on GPU nodes will be a plus
  b. Vector Database, relational database, non-relational databases 
  c. Chunking strategies, embedding and ranking models, data preprocessing.
  d. prompt management
  e. tracing and monitoring
  f. evaluation frameworks and libraries, guardrails, and security.
  g. Orchestration frameworks - Haystack/Langchain/Llama Index or Lang graph/Autogen/crew.ai.
  h. Understanding of LLM model, embedding model training/tuning will be a plus

8. Experience with MLOPs technology stack and have designed one or many productions deployed solutions

   a. Scalable training using training libraries and frameworks such as TensorFlow, PyTorch, scikit-learn, Apache Spark, Dask, Ray  etc.
  b. ML orchestration with Argo, Airflow, Kubeflow or similar orchestration
  c. ML Model serving with Kserve, Seldon or similar open source engine (real-time and batch inference)
  d. ML experimentation tracking (MLFlow or equivalent)
  e. Integration Data storage (OpenSearch/Elastic/Postgres or feature store solutions) and Data Preprocessing pipelines (Spark, Kafka, Object stores etc.)
  f. Model evaluation and performance metric calculation
  g. Experience with ML explainability, drift detection libraries and solutions will be a plus

9. Have real project experience implementing E2E ML solutions around use cases such as anomaly detection, forecasting, recommendation, clustering etc. using supervised, unsupervised and deep learning technique.

   - Knowledge of Reinforcement learning will be a plus.

 10. Must have experience of implementing Non-functional capabilities such as Fault/Alarm management, PM Monitoring, Configuration management, Logging , Tracing, Authentication, Authorization (RBAC), Security over wire and at rest, backup restore, disaster recovery, high availability, rolling upgrades etc. of the designed solution.

 11. Performance Tuning, troubleshooting and optimizing Machine learning, RAG and agentic applications

 12. Proficient in cloud-native micro services design , experience of working with  Kubernetes, Docker, helm charts,.

13. Good industrial grade solution / product development experience around big data and AIML solution with high customer value

 14. Strong problem-solving skills, Proactive, Willingness to share knowledge and empower team, Excellent communication and presentation skills with the ability to think from first principles and work in a fast-paced, collaborative environment.

 15. Experience with Knowledge Graph based systems/solutions will be a plus.

 16. Design experience with AWS, GCP or Azure based MLOps/LLMOps stack will be a plus

 17. Telecom background will be a plus .