Mahammad Althaf Tappa

AI/ML Engineer with 6+ years of experience in Machine Learning, Deep Learning, Generative AI, MLOps, and AIOps in the healthcare domain. Hands-on with LLMs, RAG architectures, transformer models, prompt engineering, and vector databases (FAISS, Pinecone, Chroma). Built production NMT language models, speech-to-text systems on Microsoft Azure Cognitive Services, GTE embedding models, and document classification models in Python, TensorFlow, and PyTorch.

Expert in model deployment, model monitoring, model versioning, CI/CD automation, and cloud ML workflows across AWS, Azure, and GCP. Strong record of processing structured and unstructured data at scale on Kubernetes-based platforms with 99.9% uptime.

Senior DevOps / MLOps Engineer · AI Engineer - IQVIA

(2022-01)

Built and productionized Generative AI and Machine Learning systems for healthcare translation workflows: transformer-based NMT (neural machine translation) language models, speech-to-text pipelines on Microsoft Azure Cognitive Services (Azure AI Speech), GTE embedding models, and Deep Learning document classification models using Python, TensorFlow, PyTorch, and Hugging Face Transformers.

• Designed and deployed LLM-based RAG applications and AI agents with LangChain and LangGraph — advanced chunking and embedding strategies, prompt engineering, vector databases (FAISS, Pinecone, Chroma), guardrails, and LLM evaluation for clinical and life-sciences use cases.
• Owned end-to-end model deployment and cloud ML workflows across AWS (SageMaker, Bedrock, EKS), Azure (Azure ML, AKS, Databricks, Azure OpenAI), and GCP (Vertex AI, GKE) — supporting 50+ production ML and LLM models at 99.9% uptime.
• Implemented model versioning, model registry, and experiment tracking with MLflow; automated retraining, A/B testing, and canary/blue-green rollouts with rollback — reducing deployment cycle time by 70% through GitLab CI/CD, Docker, Kubernetes, Helm, and Terraform automation.
• Established model monitoring and full-stack observability: data and model drift detection (PSI, KS-Test, Jensen-Shannon), LLM cost and safety monitoring with Arize AI, Prometheus, Grafana, ELK, and OpenTelemetry.
• Deployed GenAI-driven AIOps for log anomaly detection, automated root-cause analysis, and self-healing workflows — reducing production MTTR by 45%.
• Optimized real-time inference serving with FastAPI, gRPC, and Kubernetes HPA tuning — cutting latency 40% and meeting sub-100 ms SLAs.
• Engineered data, feature, and embedding pipelines over structured and unstructured data (clinical documents, transcripts, multilingual text) on Azure Databricks and Snowflake with DataOps quality gates and lineage tracking.
• Embedded Responsible AI and healthcare compliance (HIPAA/GDPR): PII detection and redaction, IAM access control, model evaluation, and policy enforcement; mentored junior engineers on MLOps standards and design patterns.

DevOps Engineer - IQVIA

(2019-06 - 2021-12)

• Built CI/CD pipelines with GitLab CI/CD, Jenkins, and GitHub Actions for containerized microservices and ML workloads on Docker and Kubernetes (AKS/EKS), enabling reliable, repeatable releases.
• Automated cloud infrastructure provisioning on AWS and Azure with Terraform (IaC) — clusters, networking, IAM, and Key Vault — reducing manual provisioning effort by 80%.
• Implemented monitoring, alerting, and log analytics with Prometheus, Grafana, ELK, and Azure Monitor; drove incident management, RCA, and SLO/SLA adherence for production healthcare applications.
• Developed Python and Bash automation for deployments, patching, database schema migrations (Azure SQL, Snowflake), and routine operations — cutting incident response time and manual toil.
• Administered Linux servers and Kubernetes clusters across Dev, QA, and Production with strict change management and high-availability uptime targets.

M.Tech - Electronics & Telecommunication - JNTU Anantapur (2013)

B.Tech - Electronics & Communication Engineering - JNTU Hyderabad (2010)