• Open to AI/ML roles
  • Relocation flexible

Adarsh Pala

Master's Student — Indian Institute of Science, Bengaluru

I build machine learning &
computer vision systems.

Expertise

  • Computer Vision
  • LLMs & RAG
  • Agentic AI
  • Analytics

Spanning the full ML lifecycle — from problem framing and model training to evaluation, monitoring and measurable product impact.

View my work Download CV

Experience

Work Experience

Why each role matters, what it maps to, and the measurable signal it produced.

  • 3Years industry
  • 0.97mAP@0.5 detection
  • 97.6%count-level agreement
  • Atlas Copco
  • Xylem
  • IISc Bengaluru

AI/ML Intern

Atlas Copco

India

01-May-2026 to 30-Jul-2026

Built AI systems that turn complex engineering drawings and CAD knowledge into structured, searchable intelligence.

  • ↗ 0.97 mAP@0.5
  • ↗ 0.87 mAP@0.5:0.95
  • Object Detection
  • OCR
  • Agentic AI

P&ID Digitisation

End-to-end pipeline: symbol detection → OCR → graph extraction

Trained YOLOv11s to detect symbols on large industrial diagrams (mAP 0.97), then used PaddleOCR to read instrument tags and extract connectivity graphs automatically.

  • YOLOv11s
  • PaddleOCR
  • Python

Senior Engineer

Xylem Water Solutions

India

Apr 2022 – Apr 2025

Owned technical pre-sales for water-infrastructure projects — from customer requirements to winning proposals.

  • ↗ 3 years
  • Technical proposals
  • Tender engineering

Customer engineering

Pre-bid & post-bid tender support

Worked with sales managers, distributors, key accounts and end customers across the full bidding lifecycle.

Proposals

Specification-driven selection

Selected pumps against customer specifications and prepared technical proposals, clarifications and deviation lists.

About

A bit about me

I'm a machine-learning practitioner pursuing my Master's at the Indian Institute of Science, Bengaluru. I like owning problems end to end — framing them, building the systems that solve them, and making the results measurable.

Currently
Master of Business Analytics & Technology Management — Indian Institute of Science, Bengaluru Class of 2027
Previously
Three years as Senior Engineer at Xylem Water Solutions · B.Tech, Nirma University
Focus
Machine learning, analytics and AI-driven products that create measurable impact
Certified
AWS Machine Learning Engineer — Associate

Projects

Featured Projects

Macro photograph of engineering fasteners
Agentic AI · InternshipMulti-agent + MCP

Multi-Agent CAD Copilot — Autodesk Inventor

LangGraph multi-agent system — supervisor, specialist tool-calling workers and a reflection-based reviewer — over a Chroma + Azure OpenAI RAG base for fastener lookup/generation and VAVE cost analysis. MCP tool server, per-step tracing with token & cost tracking, GPT-4o-mini fallback, and deterministic guardrails against hallucinated engineering values.

  • LangGraph
  • MCP
  • Chroma
  • Azure OpenAI
Ultrasound scan showing a fetal skull cross-section
Deep learning research0.88 Dice · 3.1 px MAE

Fetal Skull Landmark Detection & Segmentation

Detecting fetal skull landmarks and segmenting the cranium from noisy ultrasound where no off-the-shelf model existed — custom CNN regression + U-Net segmentation with ellipse fitting, experiments across three model variants, and clinically meaningful evaluation.

  • PyTorch
  • U-Net
  • CNN
  • Medical imaging
Tennis player mid-swing with neon light trails on a dark court
Computer vision · self-initiated97.6% agreement

SwingLoop — Tennis Serve Analysis Pipeline

End-to-end pipeline for tennis serve analysis — TrackNetV2 ball tracking, MediaPipe/ViTPose pose estimation and trajectory + audio + pose fusion to find the exact contact frame, served as a FastAPI microservice.

  • PyTorch
  • TrackNetV2
  • MediaPipe
  • FastAPI
View on GitHub
Industrial piping and valves in a process plant
Industry internshipmAP@0.5 = 0.97

P&ID Diagram Digitisation

Trained YOLOv11s to detect symbols on large industrial diagrams (mAP 0.97), then used PaddleOCR to read instrument tags and extract connectivity graphs — built at Atlas Copco.

  • YOLOv11s
  • PaddleOCR
  • ISA S5.1
CNC machining centre milling a metal part
Peer-reviewed publicationup to 97% accuracy

Machine Vision for Tool Life Prediction in Turning

Co-authored paper in The International Journal of Advanced Manufacturing Technology — gradient-boosted trees and support vector regression on machine-vision features (MATLAB) for tool life prediction on real manufacturing data.

  • MATLAB
  • Machine vision
  • GBT / SVR
Read the paper
GitHub repository preview for media-bias-detection
NLP · TransformersRoBERTa

Media Bias Detection in News Headlines

Fine-tuned RoBERTa transformer classifying news headlines as left, right or neutral.

  • RoBERTa
  • Python
View on GitHub

Skills

What I work with

Core

  • Python
  • MATLAB
  • Linux
  • Git

Used in

  • SwingLoopPython computer-vision pipeline end to end
  • Tool Life PredictionMATLAB machine-vision models for the published paper

Deep Learning & CV

  • PyTorch
  • CNNs
  • U-Net
  • Transformers
  • YOLOv11
  • TrackNetV2
  • MediaPipe / ViTPose
  • PaddleOCR
  • Image processing

Used in

  • Fetal Skull Researchcustom CNN regression + U-Net segmentation, 0.88 Dice
  • P&ID Digitisationtiled YOLOv11s + PaddleOCR at mAP@0.5 = 0.97
  • SwingLoopTrackNetV2 tracking + MediaPipe pose, 97.6% agreement

ML Engineering & Agents

  • AWS (certified)
  • LangGraph
  • LangChain
  • MCP
  • RAG
  • GPU training pipelines

Used in

  • CAD CopilotLangGraph multi-agent system with MCP tools over RAG
  • Vision projectsGPU training pipelines across all model work

Analytics & Management

  • Business analytics
  • Gradient-boosted trees
  • Support vector regression

Used in

  • Tool Life PredictionGBT & SVR models, up to 97% accuracy
  • Bank Term DepositXGBoost + SQL feature engineering, 89% accuracy

Education

Education & certifications

2025 — 2027

CGPA 7.9

Indian Institute of Science, Bengaluru

Master of Management — Business Analytics & Technology Management

  • Specialization: Business Analytics & Technology Management
  • Relevant coursework: Machine learning, applied statistics, optimisation, and technology strategy

2017 — 2021

CGPA 7.37

Nirma University

B.Tech — Mechanical Engineering

  • Relevant coursework: Engineering mathematics, computer-aided design, and numerical methods

Certifications

Contact

Let's connect

I'm always interested in discussing machine learning and computer vision — roles, research collaborations, or just connecting with fellow engineers and researchers.

  • Bengaluru, India
  • IISc — Class of 2027