Sankalp Jr.
0%
Sankalp Jr.
tap anywhere to enter
About Experience Projects Skills Playground Contact
Bengaluru, India Associate Manager, AI Engineering · Tredence

Sankalp Jr.

Building LLM applications, RAG pipelines & agentic workflows — on a backend engineering foundation. Python · Django · LLMs · RAG · Document Intelligence

Scroll to explore
LLM Application Development
Retrieval-Augmented Generation
Agentic Workflows
Document Intelligence
Python · Django
Backend Engineering
01 / About

I take GenAI the last mile — from prototype to production.

Associate Manager, Data Science at Tredence, focused on AI engineering — LLM applications, RAG pipelines, and agentic workflows. Behind that, 5+ years building scalable backend systems, APIs, and AI-powered workflow solutions with Python and Django.

0
Years experience
0
Years at GALE
2nd
Hackathon place
Problems solved

I specialize in backend engineering, workflow automation, and LLM-enabled applications — including role-aware assistants, RAG-based workflows, and document intelligence systems.

Previously at GALE, where I contributed to enterprise products including Everest and LedgerAI — improving performance, modernizing legacy systems, and automating manual processes with AI.

The model is rarely the hard part. Retrieval quality, access control, orchestration, and evaluation are what decide whether GenAI makes it into production.
Agentic Workflows RAG Systems LLM Engineering Document Intelligence API Design Cricket 🏏
02 / Experience

Backend discipline, applied to AI products.

Jul 2026 — Present
Tredence Inc.Bengaluru, India
Associate Manager, Data Science Current
  • Building LLM applications, RAG pipelines, and agentic workflows for enterprise AI solutions.
  • Taking GenAI the last mile into production — spanning retrieval, document intelligence, and Python/Django backends.
PythonLLMsRAGAgentic WorkflowsDocument Intelligence
Apr 2024 — Jul 2026
GALEBengaluru, India
Senior Associate, Core Technology
  • Owned backend feature delivery and release execution across AI-led products.
  • Contributed to LedgerAI — a document-intelligence platform that extracts structured data and automates downstream workflows.
  • Mentored junior engineers, reviewed implementations, and raised delivery quality through stronger task ownership.
  • Partnered with product and business stakeholders to ship scalable, user-centric solutions.
PythonDjangoDRFPostgreSQLCeleryRedisLLMs
Apr 2022 — Apr 2024
GALEBengaluru, India
Associate, Core Technology
  • Delivered major releases for Everest — refactoring legacy code, optimizing APIs, and streamlining backend workflows.
  • Built an LLM-enabled assistant that answered questions and executed role-based actions via natural language, including CRUD gated by access level.
  • Drove early AI adoption through conversational interfaces, enterprise access control, and workflow automation.
  • Won 2nd place in the GALE internal hackathon.
PythonDjangoREST APIsLLMsPostgreSQL
Jan 2021 — Apr 2022
GALEBengaluru, India
Junior Associate, Core Technology
  • Contributed across frontend and backend for Everest, building hands-on Django web-development experience.
  • Supported feature development, bug fixes, and improvements to usability, stability, and release readiness.
  • Proactively improved code quality and platform reliability beyond assigned tasks.
PythonDjangoJavaScriptHTML/CSS
03 / Selected Work

Where the hard part wasn't the model.

Enterprise products at GALE — publicly named on my profile.

◆ 01 — Flagship
LedgerAI
AI-powered document intelligence · Backend / AI engineering
Extracts structured data from unstructured documents and automates the workflows that follow — extraction pipelines, downstream orchestration, and the plumbing that decides whether document AI is actually reliable in production.
Document IntelligenceExtraction PipelinesWorkflow AutomationOrchestration
▲ 02
Role-Aware LLM Assistant
Everest · Builder
Answers questions in natural language and executes role-based actions, including CRUD gated by the user's access level. Built before "put an LLM in the product" was a roadmap item — the hard problem was making an assistant safe to give write access.
Conversational UIAccess ControlNL-to-ActionPermission-Aware
● 03
Everest
Enterprise platform · Full-stack → backend
A multi-year enterprise product I worked across all three GALE roles — legacy modernization, API optimization, performance refactoring, and reliable release delivery.
Legacy ModernizationAPI OptimizationPerformanceRelease Delivery
04 / Skills

What I work with.

Generative AI
LLM & AI Engineering
LLM App Development RAG Agentic Workflows Prompt Engineering Vector Databases Document Intelligence
Core Stack
Backend & APIs
Python Django Django REST Framework REST API Design API Optimization
Infrastructure
Data & Orchestration
PostgreSQL Celery Redis Apache Airflow Airflow DAGs
Practices
Architecture & Craft
Low-Level Design High-Level Design Git Workflow Automation Spec-Driven Dev
Focus Areas
Pinned
LLM Application Development Retrieval-Augmented Generation Python
Ways of Working
Toolkit
AI-Assisted Prototyping Rapid Iteration Code Review Mentoring
05 / Education

Academic foundation.

Postgraduate
Master of Computer Applications
MCA · Computer Engineering · Sep 2018 – Dec 2020
Dr. Ambedkar Institute of Technology
Bengaluru, India
Undergraduate
Bachelor's Degree
2015 – 2018
Surana College
South End Circle, Bengaluru
06 / Playground

Take a break.

Four mini-games. One built to work your memory.

01 — N-Back · working memory
A tile lights up each round. Hit Match when the position is the same as N steps ago. The real cognitive-science memory test.
N =
Press Start. Watch the tiles.
Trial: 0/20
Hits: 0
Accuracy:
02 — Code Breaker
Crack a 4-symbol secret code in 8 attempts. ● = right spot. ○ = wrong spot.
Press Start to play.
Attempts: 0/8
03 — Type Speed
How fast can you type? 30 seconds. Hit Start and go.
Press Start to begin.
WPM:
Accuracy:
Time: 30s
04 — Tic-Tac-Toe · vs unbeatable AI
You are X. Play vs a minimax AI. Good luck drawing.
Press Start to play.
You (X): 0
AI (O): 0
Draws: 0
07 / Contact

Let's build
something.

Open to interesting problems in AI engineering — LLM applications, RAG, and the unglamorous parts that decide whether GenAI ships.