Akshay Bhardwaj
Director of Data Science and Ai @Mechademy
Signup · Get unlimited contacts
WORK HISTORY
Director of Data Science and Ai @Mechademy
New Delhi, IN
15+ & - Directing 10 data scientists and ML engineers, scaling to 24+ with strategic hires across GenAI, MLOps, Data Engineering, and Frontend AI. Defined and driving 7 major 2026 AI initiatives aligned with CEO vision: Diagnostic Copilot, Agentic Monitoring, ML Democratization, Unified Data Lake (50TB+), and more- Built a LangGraph-based Diagnostic System automating fault discovery, sensor mapping, and root cause analysis — turning weeks of expert work into hours- Architected a 50TB+ Data Lake with Medallion Architecture achieving 97% query performance improvement and 98% cost savings vs. traditional warehouses- Deploying production GenAI across all clients: VLM-powered Model Validation for deployment readiness and optimal thresholds, plus LLM-driven Feature Selection- Pioneered classification ML model creation with Synthetic data - Expanding the ML portfolio — VAE Autoencoders, TabNet, Deep Isolation Forest, Neural Anomaly Detection — building holistic equipment health views for IoT time-series- Built Automated Curve Digitization via Computer Vision (fine-tuned segmentation + image processing), replacing hours of manual work. Automated configuration from datasheets using LLMs, eliminating manual data entry per onboarding- Evolved our AutoML to production excellence: model creation from 2 months → 3 days → 50 models/day. One person deploys 20 models daily; 80% production-ready out of the box — driven by Automated Feature Selection, Validation, and Parallelized HPO. Reduced minimum data requirement from 2 years to ~3 months while improving model quality.Key clients: Chevron, Berkshire Hathway, Woodside LNG, INPEX, etc
EDUCATION
Udacity
Machine Learning Advanced Nanodegree, Machine Learning
Udacity
Data Engineering Nanodegree, Data Engineering
Udacity
Machine Learning Foundation Nanodegree, Machine Learning
Udacity
Machine Learning Basic Nanodegree, Machine Learning
Udacity
Cloud DevOps Engineer, Cloud Technologies
Maharashtra Institute of Technology
Bachelor's Degree, Petroleum Engineering
ABOUT AKSHAY BHARDWAJ
I lead Data Science & AI at Mechademy, where we\'re building the AI backbone behind Turbomechanica — an Industrial IoT platform that monitors and predicts the health of critical turbomachinery for companies like Chevron, Woodside, and Devon Energy.Over 7 years, I\'ve grown from the team\'s first data scientist to Director — wearing every hat along the way: DevOps, data engineering, ML research, platform architecture, and now strategic leadership. That\'s startup life, and I wouldn\'t trade it.What we\'ve built- A domain-specific AutoML platform on Ray: our first regression model took 2 months to build. Today, one person creates and deploys 20 models in a day — and 80% need zero manual tuning- Production GenAI integrated into our ML lifecycle — automated feature selection, model validation, and deployment recommendations running live across enterprise clients. Not prototypes- ML serving that evolved from 1 model per K8s pod to 2 EC2 instances serving 400+ models — a journey I architected end-to-end- Orchestration evolution: Airflow → Dagster → now Celery + Ray, with our server as the orchestrator- A 50TB+ enterprise data lake (Apache Iceberg, medallion architecture) with 97% query speedup and 98% cost savings vs. Snowflake- Cloud-agnostic infrastructure: deployed clients simultaneously on AWS and Azure before consolidatingCurrently driving 7 major AI initiatives for 2026: a Diagnostic Copilot that automates fault discovery and root cause analysis, Agentic Monitoring to scale equipment coverage per engineer by 10x, ML democratization so anyone can build models, and a unified data lake at 50TB+ scale.I\'m actively hiring exceptional people who want to work at the intersection of GenAI, distributed systems, and industrial-scale ML. If that sounds like you — or someone you know — let\'s talk.Background: B.E. in Petroleum Engineering (MIT Pune) — the domain expertise that lets me build physics-informed AI for the energy sector. CKAD certified. Self-taught in ML through Udacity nanodegrees, fast.ai, and relentless side projects — from building transformers from scratch to multi-agent systems with LangGraph.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.