room records mapped with AI-assisted ranking pipelines
Backend, Data & AI Systems Engineer
I build production systems that make search, data, and developer workflows faster and more reliable.
I'm an SDE-2 currently working on large-scale travel infrastructure at TripJack, with previous experience across backend, data, search, and automation at Trademo.
hotel search latency improvement from 60s to under 200ms
pages indexed through optimized high-traffic APIs
automated data pipelines shipped with Python and Airflow
About
I like owning systems end-to-end, from database internals up to the API contract.
Most of my work sits where performance, reliability, and applied AI meet: rewriting slow search paths, building the plumbing that lets 50+ suppliers talk to one platform, and giving teams AI tooling that saves them from digging through logs by hand.
What I look for in a problem is whether the fix compounds - a 300x latency win, a supplier onboarding flow that no longer needs a human in the loop, an MCP server that turns a 20-minute debugging session into a two-line query. I'd rather ship the boring, durable version of something than the clever one that needs babysitting.
Experience
Recent roles and the systems behind them
TripJack
SDE-2
- Built a real-time hotel room mapping pipeline using FAISS, Milvus, Ollama-powered LLMs, and neural reranking to map 100K+ room records with about 95% top-1 accuracy.
- Developed a Golang microservice with gRPC integrations for 50+ suppliers, cutting onboarding time from more than 4 weeks to under 1 week.
- Optimized hotel search using BitSet filtering, caching, and better pagination, reducing response time from 60s to under 200ms.
- Added distributed observability with metrics, tracing, and log correlation, reducing debugging time by 60%.
- Built internal AI tooling, including MCP servers across PostgreSQL, MongoDB, and Elasticsearch, plus an AI-powered PR review workflow.
Trademo
SDE-1, Backend and Data
- Led migration from a Python monolith to Golang microservices, improving scalability and reducing latency by 50%.
- Architected an ACL system with dynamic, role-based permission management for modular product features.
- Designed a retrieval-augmented classification workflow for HS codes using LangChain and vector search.
- Built 30+ automated data pipelines with Python and Apache Airflow and improved bulk workloads with RabbitMQ and Celery.
- Optimized Elasticsearch indices, mappings, and queries, and improved high-traffic APIs that helped index more than 8 million pages.
Earlier internships
Python, data engineering, and backend foundations
- Artyvis Technologies: web scraping and ETL automation across 50+ websites and 10+ pipelines.
- S759Labs Technology LLP: extracted financial data from company balance sheet PDFs using Python-based parsing workflows.
- TalentCruz: worked on Django backend features, database tasks, frontend integration, and Razorpay payment APIs.
Selected Highlights
Work that shows how I think
Search & Matching
AI-assisted room mapping at travel scale
Combined vector search, reranking, and LLM reasoning to automate supplier room matching and improve accuracy on high-volume catalog data.
Platform Engineering
Microservices and orchestration that reduced bottlenecks
Built and tuned Golang and Java services handling large request volumes, with stronger routing, validation, failover, and observability patterns.
Developer Productivity
Internal AI tooling for data access and PR review
Shipped MCP-based tooling that lets teams query production-shaped data safely and accelerate code review feedback loops.
Projects
Something I built end-to-end, outside of work
Personal Project
PostgreSQL Query Monitoring & Analysis Platform
Built a PostgreSQL extension that captures query execution metrics through server hooks, buffers events, and streams them to a gRPC service for analysis. The result is a lightweight monitoring layer for slow query detection, workload visibility, and deeper database performance insights.
C · Golang · gRPC · PostgreSQL internals
Skills
Tools I reach for often
Languages
Python, Golang, JavaScript, C++, C
Frameworks & Runtime
Django, Flask, FastAPI, Gin, Celery, gRPC, RabbitMQ, Kafka, Apache Airflow
Cloud & DevOps
AWS, GCP, Docker, Kubernetes, ArgoCD, Jenkins, New Relic, Sentry
Data & Search
PostgreSQL, MySQL, MongoDB, Redis, Elasticsearch, Milvus, Aerospike
Testing & Automation
Selenium, Scrapy, Playwright, ETL workflows, scraping pipelines
Applied AI
RAG systems, vector search, LangChain workflows, LLM-assisted tooling, neural reranking
Education
ABES Institute of Technology
Bachelor of Engineering in Computer Science and Engineering
Ghaziabad, India - August 2018 - June 2022 - Aggregate: 75.2%
Recognition
- Grand Finalist, Smart India Hackathon 2020
- Winner, Internal Hackathon at ABESIT for SIH 2020
- Coursework across Deep Learning, Computer Vision, NLP, DBMS, and Software Engineering