Background
Senior MLOps & AI Platform Engineer

Engineering Production-Grade MLOps & AI Platforms

Bridging the gap between Data Science innovation and Production reliability. Deployed on Azure, Databricks, and Edge AI. Focused on LLMOps, Model Governance, and Cost Optimization.

From Notebooks to Scalable Platforms

I specialize in building scalable, secure, and self-healing AI platforms. I help enterprises navigate the complex transition from experimental Jupyter notebooks to robust, production-grade cloud environments.

My approach focuses on creating automated governance and robust monitoring systems that empower Data Scientists rather than slowing them down. I believe that MLOps isn't just about automation—it's about building trust in AI.

12+ Years
Total Experience
3.5+ Years
MLOps Specialization
Azure & Databricks
Cloud Focus
Self-healing Systems
Reliability
Enterprise Ready

Deploying LLMs and Computer Vision models to edge and cloud with zero-downtime architectures.

Mastered Technologies

Azure
AWS
Kubernetes
Docker
Terraform
Databricks
MLflow
Azure ML
Apache Spark
Apache Airflow
LangChain
LangGraph
CrewAI
OpenAI
OpenVINO
ONNX
Hugging Face
Dynatrace
Grafana
Prometheus
Snyk
GitHub Actions
Azure DevOps
Jenkins
Python
Rust
Bash
Azure
AWS
Kubernetes
Docker
Terraform
Databricks
MLflow
Azure ML
Apache Spark
Apache Airflow
LangChain
LangGraph
CrewAI
OpenAI
OpenVINO
ONNX
Hugging Face
Dynatrace
Grafana
Prometheus
Snyk
GitHub Actions
Azure DevOps
Jenkins
Python
Rust
Bash

Professional Experience

A decade of engineering excellence across the UK and India.

Sep 2022 – Present

Senior MLOps Engineer

TATA CONSULTANCY SERVICESLondon, UK

Delivering enterprise ML infrastructure for Marks & Spencer (top UK retailer).

  • Architected an automated deployment validation system that replaced manual multi-team sign-off with policy-based checks, cutting deployment lead time 83% (from 2 hours to 20 minutes).
  • Managed an end-to-end ML platform on Azure, using Databricks for scalable model training and Azure Kubernetes Service for high-availability inference serving.
  • Consolidated fragmented, duplicate feature tables into a centralized feature store, eliminating data silos and speeding up the feature engineering lifecycle by 40%.
  • Enforced compliance and governance standards, including GDPR adherence, across the data science lifecycle for sensitive retail data.
  • Designed operational dashboards to monitor model drift and system health, giving real-time visibility to both business and operations stakeholders.
  • Built custom plugins for Apache Airflow and Databricks Operators to extend pipeline orchestration capability.
  • Maintained a shared feature-engineering repository and coordinated requirements across multiple enterprise data science teams.
DatabricksAirflowMLflowAzureDynatrace
Aug 2021 – Aug 2022

Data Scientist

DAVE.AIBangalore, India

Focused on Edge AI optimization and cost-effective infrastructure.

  • Partnered with Intel to optimize Automatic Speech Recognition (ASR) and NLP models using the OpenVINO framework, deploying them on edge devices for Quick Service Restaurant (QSR) applications.
  • Re-architected inference infrastructure for QSR clients using optimized inference engines, cutting per-store infrastructure cost by 40%.
  • Built ASR models achieving a Word Error Rate of 0.1–0.2 across Banking, Retail, Food Chain, and Automotive domains.
  • Worked as an end-to-end ML engineer on a food-chain domain project spanning image, speech, and text analysis.
  • Built a benchmarking framework to evaluate hardware requirements for in-house AI components.
  • Partnered with Nvidia on a proof-of-concept using Jetson devices for the QSR system.
PythonTensorFlowNLTKOpenCVOpenVINOAWSGCP
Apr 2016 – Aug 2021

Business Analyst

ANSRSOURCEBangalore, India

Led product strategy, operations analytics, and QA automation.

  • Built and led a team of 70 from the ground up, generating $1.5M in Annual Recurring Revenue.
  • Developed an in-house QA automation tool that cut manual data-categorization effort by 60% and raised data accuracy to 99.5%.
  • Used predictive models to shape strategies addressing growth and operations issues.
  • Defined and tracked KPIs for existing and new functions/projects across the company.
  • Partnered with 20+ data analysts on data quality, integrity, and new approaches to presenting existing data.
  • Acted as liaison across operations, technology, vendors, and clients, managing BAU support alongside project work.
  • Collaborated with stakeholders to refine product requirements, functional specs, and product direction.
PandasJIRATableauExcel VBAPower BIIBM Watson StudioMySQLScikit-Learn
June 2014 – Apr 2016

Content Programmer

ANSRSOURCEBangalore, India

Foundation in software engineering and web platforms.

  • Developed interactive e-learning platforms for publishers and universities using Python, Django, and Flask.
PythonDjangoFlaskMySQLJavaScriptMongoDBReact

Academic Background

The theoretical foundations of my engineering practices.

2026 (In Progress)

AI Performance Engineer Fellowship

Nebius Academy

Advanced fellowship program focusing on high-performance training, inference, and hardware-aware optimization for LLM models.

2021 – 2022

Post Graduate Program – AI & ML

Great Learning (University of Texas, Austin)

Specialized training in Artificial Intelligence and Machine Learning, focusing on deep learning, computer vision, and NLP.

2012 – 2014

Master of Science (Mathematics)

Bangalore University

Advanced mathematics degree providing the theoretical and statistical foundation for machine learning algorithms and data science.

2009 – 2012

Bachelor of Science (Stats, Math, CS)

Bangalore University

Undergraduate degree in Statistics, Mathematics, and Computer Science, establishing strong computational and statistical foundations.

Featured Projects

Deep engineering dive into automated governance, LLMOps, telemetry, and edge intelligence.

🚦 Traffic Light Governance

Automated Model Governance as Code. A GitHub Actions "Gatekeeper" that enforces a quality gate on every PR — validating code quality, security posture, and model performance regressions before a single line reaches production.

PythonGitHub ActionsPyTestSnykSemgrep

🛠️ Airflow Custom Plugins

Airflow Custom Plugins for seamless integration with various services. End-to-end LLM pipeline with automated evaluation loops — GPT-4 acts as a judge to score prompt quality, enabling data-driven prompt engineering at scale with full MLflow lineage tracking.

AirflowCustom PluginsPythonAzure

👁️ Retail-Lens

AI-Powered Smart Shelf Vision. Computer vision system that empowers store associates to instantly identify out-of-stock items, misplaced products, and incorrect price tags — reducing shelf compliance issues in real time.

Google LensOpenCVDockerPythonEdge AI

🧠 AI-Powered QSR — Edge Compliance & Benchmarking

A benchmarking study of small form factor hardware devices that manage AI workloads deployed at the edge.

Intel OpenVINOOpenCVPythonEdge HardwareAWSASRNLP

⚡ Skywalker: Automated QA using Classification

A fully automated QA analysis engine using Classification & Clustering

Scikit-learnPythonpandasnumpy

🛠️ Serverless Model Profiler

CloudNative Hackathon submission. Developed a serverless benchmarking suite that automatically spins up isolated model inference tasks in AWS Lambda containers to profile model latency, cold-start, and memory drift.

AWS LambdaDockerPythonMLflow

🧞 AdGenie

LLM Lifecycle Management with "Prompts as Code". End-to-end LLM pipeline with automated evaluation loops — GPT-4 acts as a judge to score prompt quality, enabling data-driven prompt engineering at scale with full MLflow lineage tracking.

LangChainMLflowOpenAIAzurePython

Engineering Philosophy

Principles that guide my work in high-stakes AI production environments.

Zero-Friction for Data Scientists

Abstract away Kubernetes, Docker, and infrastructure entirely. Data Scientists should think in models and mathematics — not YAML configuration files.

Continuous Automation

From latency profiling to security scans and validation gates. If a task is performed twice, it must be codified into a reusable operator, plugin, or CI/CD runner.

Structure Begets Speed

Ad-hoc scripts don't scale to 50+ models in production. Enforcing strict project templates and CI/CD contracts makes every deployment repeatable, audit-ready, and automated.

Guardrails Enable Confidence

Strict governance (like the Traffic Light validation system) lets teams deploy faster — not slower — because safety is baked in, not bolted on.

Frugal Architecture

Cloud costs should not grow linearly with model usage. Optimised inference (ONNX, quantization, edge deployment) is not a nice-to-have; it's a first-class engineering concern.