Pratik Kumar
Senior Software Engineer @ Elucidata
Building AI systems that handle
complex data at enterprise scale.
I build and scale production systems across backend, distributed infrastructure, and AI, turning complex data and workflows into reliable, user-facing platforms.
5
Years of Experience
Monsanto & Elucidata
6+
Organisations Served
Elucidata
5+
Products Shipped
Elucidata
Timeline
Proposed & drove Neptune→Neo4j migration · 55% latency reduction · grew to 3 enterprise clients
Built Polly KG platform (React + Django + AWS) · first AI layer: NL query system on LangChain + OpenAI
Built data visualization platform end-to-end · React + Express + Plotly · 10+ omics repos
Large-scale single-cell analysis on 2GB+ datasets · ScanPy + AWS (S3, EFS, ECR)
Analytics tool in Python + Plotly + Shiny · GCT files over 1GB · converted to full-time
Sales data visualization web app · Python + Dash + Plotly
Tech Stack
Languages
Python, JS
AI / LLM
LangChain, LlamaIndex, LangGraph
Backend
Django, FastAPI, Flask
Frontend
React, Next.js
Data
Neo4j, Postgres, SQLite, MongoDB
Cloud
AWS, Docker, CircleCI
Selected Work
Polly KG 2025
Built Polly KG from scratch on AWS Neptune. Proposed and drove the re-architecture to Neo4j. Designed multi-agent orchestration across multiple large-scale knowledge graphs. Grew platform from 1 to 3 enterprise clients. Own the full platform roadmap.
LangChain · LlamaIndex · Neo4j · AWS · FastAPI
Polly Co-Scientist 2024
Designed and shipped the NL query system over biomedical graphs. Graph-RAG pipelines with evaluation reporting.
LangChain · OpenAI · React · Django · AWS
Gene Exploration Platform 2022
Built visualization platform serving 10+ omics repositories. 50% faster data prep via Pandas pipeline. Scaled ETL to handle 8+ public data sources.
React · MongoDB · Plotly · AWS
Open Source
GitHub →TermiPy
Interactive shell for Python with auto‑completion, syntax highlighting, and plugin support
rlimit
Lightweight rate‑limiting library for Python APIs with configurable strategies and clean integration patterns
hscale
Scaling and load‑balancing toolkit for distributed services with dynamic routing, health checks, and auto‑scaling signals
Newsletter
Deep dives on AI & engineering
Monthly essays on LLM systems, knowledge graphs, and lessons from shipping production ML.
Subscribe free →Chronicles
All posts →
Through the Lens
view all
Let's work together
Open to discussing AI/ML projects, engineering leadership roles, or just chatting about knowledge graphs.