AI Engineer & Team Lead — The Unchain
Harsh Pipalia
I lead a team of four engineers at The Unchain building Neuron — an AI knowledge platform that turns a team’s scattered conversations into decisions, risks, and answers with sources. MS in Data Science from Lehigh University; based in Bethlehem, PA.
Bethlehem, PA · harshpipalia143@gmail.com · +1 (484) 635-5214
Experience
Where I’ve worked
Nov 2025 — Present
Remote · New York
AI Engineer & Team Lead
The Unchain, Inc.
Lead a team of 4 engineers building Neuron, a multi-tenant RAG knowledge platform that ingests 24 data sources and answers questions with inline citations — owning sprint planning, code review, and architecture.
Built the retrieval pipeline: Anthropic-style Contextual Retrieval with prompt caching, reranking, and compression over pgvector embeddings.
Designed LLM signal extraction that mines decisions, risks, and action items from team communications, with PMBOK risk scoring and cross-source clustering.
Architected a 16-table knowledge ontology with immutable audit logs, cross-tool identity resolution, and human-in-the-loop proposal review, served via GraphQL.
Ship production infrastructure: Terraform on AWS, zero-downtime CI/CD, Stripe seat-based billing, Clerk multi-tenant RBAC, and a multi-provider LLM layer (Anthropic, Bedrock, OpenAI).
Apr 2024 — Jun 2024
India
Data Analyst Intern
Freebird App Studio LLP
Extracted and analyzed large gaming datasets with SQL and Python to uncover actionable insights.
Ran A/B tests and statistical models to optimize in-game features and player behavior.
Delivered insights that improved engagement and ~$15k in revenue potential.
Nov 2023 — Feb 2024
India
Business Analyst Intern
Star Techno Mech PVT LTD
Analyzed 5,000+ sales transactions to surface trends and optimize business strategy.
Applied financial modeling and data visualization for sales forecasting.
Identified growth opportunities exceeding 10%, driving strategic sales adjustments.
Featured work
Neuron — AI knowledge platform for teams
Neuron connects a company’s tools and turns everything the team says and writes into structured knowledge: ask questions and get answers with sources, surface decisions and risks automatically, and let configurable AI agents analyze problems from multiple perspectives. I lead the engineering team and own the AI pipeline end to end.
24
Data-source integrations, from Slack and Jira to Notion and Gmail
16
Purpose-built ontology tables with immutable audit logging
RAG
Contextual retrieval with reranking, compression & citations
EN·JA
Fully bilingual product — English and Japanese
FastAPI
Next.js
TypeScript
PostgreSQL + pgvector
Anthropic Claude
AWS Bedrock
OpenAI
Terraform
Docker
Stripe
Clerk
Redis
More projects
Selected data work
Data Warehouse & SQL Analytics
Modern data warehouse over 60,000+ ERP & CRM records using Medallion Architecture in SQL Server — ETL workflows into a Star Schema, plus advanced reporting with window functions and CTEs for customer segmentation and year-over-year analysis.
View on GitHub ↗
Market Basket Analysis
Apriori and FP-Growth association-rule mining over 50,000+ transactional records to identify product associations — informing cross-selling strategy and improving customer targeting by 2%.
View project ↗
Skills
What I work with
AI / ML
RAG — retrieval, reranking, compression
pgvector
Embeddings
Multi-agent systems
Prompt engineering
Anthropic API
OpenAI API
AWS Bedrock
TensorFlow
PyTorch
Backend
Python
FastAPI
SQLAlchemy
Alembic
PostgreSQL
Redis
Frontend
TypeScript
Next.js
React
Tailwind CSS
Data
SQL
NumPy
Pandas
Scikit-Learn
Google BigQuery
Kafka
PySpark
Cloud / DevOps
AWS — EC2, RDS, Bedrock
Terraform
Docker
GitHub Actions CI/CD
Education
MS in Data Science
Lehigh University, Bethlehem, PA — Graduated December 2025 · Merit Scholarship
B.Tech in Computer Engineering
Marwadi University, Rajkot, India — Graduated April 2024
Certifications
Languages: English (fluent) · Hindi (fluent) · German (basic)
Contact
Let’s talk
Open to conversations about AI engineering, RAG systems, and building products with small teams. The fastest way to reach me is email.
LinkedIn · GitHub · +1 (484) 635-5214 · Bethlehem, PA