Vikram Anand
Principal ML Engineer / Head of AI
Bengaluru, India • +91 90080 47193 • vikram.anand@proton.me • linkedin.com/in/vikram-anand-ai • github.com/vikramanand
Portfolio: www.eliorexa.com/portfolio/ai-ml-engineer-senior
Professional Summary
AI leader and ML architect with 12 years building machine learning platforms and production systems at scale. Sets ML strategy across LLM, NLP and recommendation systems, leads teams of 10–20 engineers and scientists, and ties model investments to revenue and efficiency. Drove research-to-production programs, built the MLOps platform behind dozens of models, established model governance and responsible-AI practices, and mentors engineers into staff and lead roles.
Skills
Leadership: AI strategy & roadmap, Team building (10–20), Hiring & mentoring, Stakeholder & C-suite alignment, Model governance & responsible AI
ML architecture: LLM & RAG systems, Recommender systems, Distributed training, Feature & vector platforms, Real-time inference at scale
Stack: Python, PyTorch, TensorFlow, Hugging Face, Ray, Spark, scikit-learn
Platform & MLOps: AWS, GCP, Kubernetes, Kubeflow, MLflow, Terraform, Observability (Datadog, OpenTelemetry)
Core Competencies
AI leadership · ML architecture · Python · PyTorch · TensorFlow · LLMs · NLP · MLOps · model deployment · AWS · GCP · model governance · scale
Work Experience
Principal ML Engineer & Head of AI Platform — Lumora AI
Mar 2019 – Present
Bengaluru
- Lead a 17-person AI org (engineers + scientists) owning the ML platform behind 40+ production models serving 1.5B inferences/day.
- Set the LLM strategy and built a RAG-based copilot that drove ₹38 crore in new ARR and cut customer support cost 28% across the product suite.
- Architected a unified MLOps platform (Kubeflow, MLflow, feature store) that cut median model-to-production time from 6 weeks to 5 days for every team.
- Reduced GPU training and inference spend ~₹6 crore/year via distributed training on Ray, quantization and right-sized autoscaling on Kubernetes.
- Established model governance — bias audits, eval gates and an LLM red-teaming program — adopted as company-wide policy ahead of enterprise compliance reviews.
- Hired 12 engineers and scientists and grew 5 into staff/lead roles; instituted the research-to-production review used across the AI org.
ML Lead / Senior Data Scientist — Cobalt Retail Group
Jun 2015 – Feb 2019
Gurugram
- Led a team of 7 building demand-forecasting and pricing models (PyTorch, scikit-learn) across 1,200 stores, lifting forecast accuracy 19% and freeing ₹14 crore in working capital.
- Stood up the first production ML pipeline and CI/CD for models, taking the team from quarterly notebooks to weekly automated retraining.
Data Scientist / ML Engineer — Earlier roles — TCS Research, Mu Sigma
2013 – 2015
India
- Built NLP and classical ML models across banking and analytics products; promoted twice for delivery and technical depth.
Education
M.S. Machine Learning
2013
Indian Institute of Science (IISc), Bengaluru
Specialization in statistical learning and optimization
B.Tech Computer Science
2011
National Institute of Technology (NIT) Trichy, Tiruchirappalli
Certifications
- AWS Certified Machine Learning – Specialty — Amazon Web Services (2021)
- Google Cloud Professional ML Engineer — Google Cloud (2020)
Selected Achievements
- Drove ₹38 crore in new ARR with an LLM-powered product copilot built research-to-production in under two quarters.
- Cut ML infrastructure spend ~₹6 crore/year via distributed training, quantization and autoscaling.
- Speaker, PyData Bangalore 2024 — "Governing LLMs in production: evals, red-teaming and guardrails."
- Advises 2 AI startups on ML architecture; mentor at a women-in-ML fellowship.