Ashish
Singhal
GenAI Engineer specializing in multi-vector retrieval systems and production RAG architectures. Building AI agents that serve Fortune 500 clients.
5
AI Agents Shipped
$4M+
Revenue Delivered
4,000+
Users Served
Top 1%
Axtria Performer
/ 01
Core Expertise
Deep specializations in AI engineering, retrieval systems, and production infrastructure.
Multi-Agent Systems
16 active deploymentsArchitecting autonomous agent orchestration with CrewAI and LangChain. Built 5 production agents (document Q&A, web scraping, forecasting) deployed across 8 Fortune 500 clients.
RAG & Retrieval Architecture
92% answer accuracyDesigning multi-vector retrieval pipelines with Tantivy for keyword search, FAISS for Q&A pairs, and Chroma for semantic search. Intent classifiers route queries across 6 retrieval strategies.
Production AI Infrastructure
5,000+ monthly queriesShipping enterprise-grade AI systems on AWS Lambda, OpenSearch, and containerized microservices. Building chatbots handling 5,000+ monthly queries for 4,000+ users.
Workflow Automation
169 hrs saved / weekIntegrating n8n workflow automation with 12 reusable sub-workflows. Reduced repetitive tasks from 20 to 7 hours per week per engineer — 169 hours saved weekly across 13 engineers.
Full-Stack Engineering
8+ enterprise systemsBuilding end-to-end with Python (FastAPI, Flask, Django), React/Angular frontends, and Java Spring Boot backends. PostgreSQL, Redis, Docker, and CI/CD pipelines.
/ 02
Experience
From automating bank operations at HSBC to building AI agents for Fortune 500 clients at Axtria.
GenAI Engineer
Aug 2023 -- PresentAxtria Ingenious Insights — Noida, India
- —Developed 5 production AI agents (document Q&A, web scraping, forecasting) deployed to 8 Fortune 500 clients as part of 13-member GenAI team that delivered $4M+ revenue in 2026
- —Architected Unstructured Agent with multi-vector retrieval: Tantivy for keyword search, FAISS for Q&A pairs, Chroma for semantic search; intent classifier routes queries across 6 retrieval strategies
- —Shipped enterprise chatbot serving 4,000+ India employees with 5,000+ monthly queries; React frontend, Flask API, RAG backend on AWS Lambda + OpenSearch + FAISS; 92% answer accuracy
- —Created agent orchestration dashboard with CrewAI + LangChain — business users configure and deploy agents without code; 16 active deployments across scraping, reporting, analytics
- —Integrated n8n workflow automation, creating 12 reusable sub-workflows; reduced manual tasks from 20 to 7 hrs/week per engineer (169 hours saved weekly across team)
- —Built Marketing Content Generator using GPT-4; reduced creative brief turnaround from 5 days to 8 hours for 40+ monthly requests with 85% approval rate
Automation Engineering Intern
Jan 2023 -- Jul 2023HSBC — Gurgaon, India
- —Automated 8 manual reporting workflows using Python and Power Query, reducing weekly execution time from 64 hours to 2 hours (97% reduction) across 16 business lines
- —Developed Automated Portfolio Manager consolidating financial data across 16 entities, eliminating manual Excel work and reducing report generation time by 80%
Education
2021 -- 2023
M.Tech in Computer Science
IIIT Bangalore
Bangalore, India
2017 -- 2021
B.Tech in Information Technology
G.B. Pant University
Pantnagar, India
Recognition
Top 1% Performer
Axtria 2026 — Ranked top 40 of 4,000+ employees
GATE AIR 943
Computer Science 2021 — Top 1% of 90,000+ candidates
DSA Educator
50+ sessions on Unacademy — 4.7/5 rating, 1,000+ students
/ 03
Selected Projects
Production systems built for enterprise clients. Real impact, real scale.
Fortune 500 Client
Enterprise Chatbot Platform
Full-stack enterprise chatbot serving 4,000+ employees with 5,000+ monthly queries. React frontend, Flask API, and multi-vector RAG backend using AWS Lambda, OpenSearch, and FAISS.
Axtria Internal
Unstructured Agent Pipeline
Multi-vector retrieval system combining Tantivy keyword indexes, FAISS Q&A stores, and Chroma semantic search. Custom intent classifier routes queries across 6 retrieval strategies based on query type.
Axtria Product
Agent Orchestration Dashboard
No-code agent configuration platform built with CrewAI and LangChain. Business users deploy agents for scraping, reporting, and analytics without writing code.
Enterprise Client
Marketing Content Generator
GPT-4 powered content system with prompt-engineered templates. Reduced creative brief turnaround from 5 days to 8 hours for 40+ monthly requests with 85% brand approval rate.
HSBC
Automated Portfolio Manager
Financial data consolidation system across 16 entities. Eliminated manual Excel workflows and reduced weekly report generation from 64 hours to 2 hours across 16 business lines.
/ 04
Technical Stack
The tools and technologies I use to build production AI systems and scalable backends.
GenAI / LLM
- LangChain
- CrewAI
- RAG Pipelines
- GPT-4
- Claude
- Prompt Engineering
Vector & Search
- FAISS
- Chroma
- Tantivy
- OpenSearch
- Elasticsearch
- Embeddings
Backend
- Python
- FastAPI
- Flask
- Django
- Java / Spring Boot
- Celery
- Redis
Frontend
- React
- Angular
- TypeScript
- JavaScript
- Next.js
Databases
- PostgreSQL
- MySQL
- MongoDB
- Elasticsearch
DevOps & Infra
- Docker
- AWS (Lambda, S3, EC2)
- CI/CD
- Jenkins
- Ansible
- n8n
/ 05
Insights
Designing RAG Pipelines for Production
Lessons learned from building retrieval-augmented generation systems that scale reliably under real-world constraints.
Decorators in Python: A Deep Dive
Exploring the mechanics of decorators beyond the basics -- metaclasses, descriptor protocols, and practical patterns.
Why I Chose FastAPI Over Flask
A pragmatic comparison of async-first API frameworks for AI-heavy workloads and strict type safety requirements.
Vector Databases: Choosing the Right One
Evaluating FAISS, Chroma, Tantivy, and OpenSearch for different embedding retrieval workloads at enterprise scale.