Skip to content
Open to remote AI/ML roles and select consulting engagements

AI/ML Engineer · GenAI, RAG & AI Agents

I build production-ready AI systems that automate real business workflows.

AI/ML Engineer specialising in Generative AI, retrieval-augmented generation, AI agents, LLM applications and machine learning — designed for production from day one, not just as prototypes.

View my work
Explore AI Lab
Download resumeLet’s work together
Ahmedabad, India · IST (UTC+5:30) · overlaps US, EU & UAE hours
Dipak Bundheliya, AI/ML Engineer

Currently

AI/ML Software Engineer II

ProductSquads · Jan 2026 — Present

0.0+

Years building AI systems

Hands-on AI/ML engineering since January 2024.

0+

Production AI systems shipped

RAG assistants, AI agents, ML platforms, CV pipelines and LLM applications.

0%

ML accuracy on dynamic datasets

No-Code Dynamic ML Framework across multiple business datasets.

0%+

Faster time-to-insight

One-click ML workflows versus manual analysis.

0%

Detection accuracy

YOLOv8 shipping-label detection for parcel processing.

0

Domains worked in

Compliance, sales automation, logistics / computer vision and consumer products.

Core stackLangChainLangGraphCrewAIRAGLLM ApplicationsAgentic WorkflowsScikit-learnXGBoostTime-series ForecastingPineconeSemantic SearchYOLOv8PaddleOCRFastAPI

Selected work

Production AI systems, documented as engineering case studies.

Problem, constraints, architecture, decisions and measured outcome — not screenshots and buzzwords. Client and product names are withheld under NDA.

All case studies
RAG · Enterprise Knowledge AssistantEnterprise client — Compliance platform

RAG-Based Compliance Intelligence Assistant

Conversational access to US regulatory and safety standards for product teams.

  • Python
  • LangChain
  • RAG
  • Vector Database
  • Embedding Models
  • LLM APIs
  • FastAPI
  • +2
Read case study

System architecture

questioncontextproduct contextRegulatory documentsUS standards & safety lawsIngestion & chunkingstructure-aware splitsEmbedding modelsemantic vectorsVector storeclause index + metadataProduct teamchat interfaceSemantic retrievaltop-k clausesLLMgrounded answer + citationsLegacy systemsproduct & label dataCompliance summarymapped to label requirements
AI Agents · Sales AutomationSales automation system

Smart Reach Sales Agent

An AI agent that runs the outbound sales funnel — from lead discovery to booked meetings.

  • Python
  • LangGraph
  • DeepSeek
  • Gmail API
  • Hunter API
  • +2
Read case study
Machine Learning · AutoML PlatformInternal ML platform

No-Code Dynamic ML Framework

One-click regression, forecasting and anomaly detection for non-technical teams.

87%
Model accuracy
-60%
Time-to-insight
  • Python
  • Scikit-learn
  • XGBoost
  • TensorFlow
  • Statistics
  • +2
Read case study

What I build

Six ways I help teams ship AI that actually runs in production.

See services

AI Agents & Workflow Automation

Multi-step agents that execute real business processes — outreach, scheduling, research, coordination — with explicit state, tools and human hand-off.

Learn more

RAG & Enterprise Knowledge Assistants

Grounded assistants over your documents, regulations, policies and internal knowledge — with citations, access control and a refresh pipeline.

Learn more

Custom LLM Applications

Product features and internal tools powered by LLMs — personalisation, summarisation, classification, conversational interfaces — engineered for consistency and cost.

Learn more

Machine Learning Solutions

Forecasting, regression, anomaly detection and classification pipelines built on your data, with honest validation and simple interfaces.

Learn more

Computer Vision & OCR

Detection and document-understanding pipelines — labels, forms, parcels, invoices — that turn images into structured data.

Learn more

AI Backend & Production Deployment

The engineering that makes AI features dependable: APIs, containers, orchestration, monitoring and cost control.

Learn more

AI Lab

Interactive demonstrations of the patterns behind the case studies.

Sanitized, self-contained builds you can open and try. Prototypes are marked as such until they go live.

Enter the lab
SANDBOXPIPELINE / 04
PrototypeFeatured

RAG Playground

Upload a document, watch it get chunked and embedded, then ask questions and inspect exactly which passages grounded the answer.

  • LangChain
  • Vector DB
  • LLM
SANDBOXPIPELINE / 04
PrototypeFeatured

AI Agent Workflow

Step through a LangGraph-style agent as it plans, calls tools and recovers from failures — with the state visible at every node.

  • LangGraph
  • Tools
  • State machine
SANDBOXPIPELINE / 04
PrototypeFeatured

ML AutoPilot

Drop in a CSV and get profiling, feature selection and a validated regression, forecast or anomaly model — one click.

  • Scikit-learn
  • XGBoost
  • Statistics

Experience

From research intern to owning production AI systems end-to-end.

Full timeline

Jan 2026 — Present

AI/ML Software Engineer II · ProductSquads

Cuts manual regulatory research for product and compliance teams by turning hundreds of pages of standards into a queryable knowledge base.

Jul 2024 — Dec 2025

Junior Data Scientist · Bacancy

Reduced manual ML effort for non-technical teams and automated the full outbound sales funnel for sales teams.

Jan 2024 — Jun 2024

Data Science Intern · Bacancy

Faster parcel processing and automated tracking through detection + OCR.

Let’s build

Have an AI system that needs to move from prototype to production?

Remote-friendly, async by default, and comfortable working across US, UK, EU and UAE time zones from IST (UTC+5:30) · overlaps US, EU & UAE hours.

Start a conversation
Resume