AI Engineer · Portfolio 2025
Srishti
Sengupta
AI Engineer · LLM Systems · Agentic Pipelines
I build autonomous AI agents and RAG pipelines that run in production — not just notebooks. Grounded in 1.5 years of enterprise backend engineering, I bring the reliability mindset that most AI engineers lack.
Learning - Long way to go
My Story
From .NET
to the
AI Frontier
I didn't stumble into AI. I chose it — after watching LLMs go from a research curiosity to the most transformative technology in enterprise software, and realising I wanted to be on the side that builds with them, not just uses them.
My .NET background isn't baggage — it's my edge. I know what production systems actually need: error handling, observability, retry logic, clean APIs. Most AI engineers learn this the hard way. I already know it.
"The best AI engineers aren't the ones who know the most about transformers. They're the ones who can ship."
— Srishti's engineering philosophyThe Foundation (2022–2024)
BCA → MCA — Built rigorous CS fundamentals. Algorithms, databases, software architecture. Graduated with 85.5%.
The Craft (Aug 2024 – Jul 2025)
Graduate Engineer Trainee → Executive Dev at SMS Group — Enterprise C#/.NET, Oracle SQL, Agile delivery. Shipped real features to real users.
The Pivot (Parallel, 2024–2025)
LangChain → LangGraph → Agentic Systems — Built GH Bug Agent and RAG-Runner in production while working full-time. This wasn't a side hobby.
The Target (Now)
AI Engineer roles — LLM systems, agentic pipelines, RAG, LLMOps. Open to Bengaluru, Pune, Hyderabad, Remote. Ready to contribute from day one.
Technical Arsenal
What I Build With
AI & LLM Engineering
Languages & Frameworks
MLOps & DevOps
Databases
AI Tooling & Models
Engineering Practices
Academic Foundation
Education & Credentials
2022 – 2024
Master of Computer Applications (MCA)
Heritage Institute of Technology, Kolkata
Score: 85.5%2019 – 2022
Bachelor of Computer Applications (BCA)
Birla Institute of Technology, Mesra, Ranchi
CGPA: 8.5 / 10AWS Cloud Practitioner
LearningLangChain & LangGraph for Agentic AI
LearningDocker & Container Ops
LearningCareer Journey
Professional Experience
Aug 2025 – Present
Current
Executive Software Development
SMS Group · Kolkata, India
- Solely architected and built the complete application setup for a large-scale industrial project in the steel manufacturing domain. Responsibilities spanned requirements analysis, solution design, database abnormalities, API layer, and UI — end to end.
- Automated complex operational workflows using WPF, C#, EF & LINQ — reducing loading effort by ~40% across global business units.
- Designed and deployed RESTful & WCF APIs enabling seamless cross-system data integration.
- Engineered high-performance Oracle SQL queries cutting average query execution time by 30%.
- Delivered end-to-end features across SDLC in Agile sprints, consistently meeting milestones.
Aug 2024 – Jul 2025
Graduate Engineer Trainee
SMS Group · Kolkata, India
- Optimised legacy data pipelines & SQL query performance — 45% reduction in data processing time.
- Built application of Report Handling modules applying SOLID principles and large datasets,reducing post-release defect rate by 50% across sprints.
- Championed Agile ceremonies enabling rapid iteration and on-time milestone delivery.
Jan 2024 – Mar 2024
Automation IIntern
SMS Group · Kolkata, India
- Gained hands-on exposure to industrial automation workflows, office setup procedures, and enterprise engineering environments at a global steel technology firm.
- Assisted in commissioning and end-to-end testing of an automation project — supporting system validation, defect logging, and sign-off in a live plant environment.
Selected Work
Project Showcase
Agentic AI · DevTools
GH Bug Agent
Fully autonomous AI agent — ingests GitHub issue URLs, localises bugs via Python AST, generates patches with Groq LLaMA 3.3-70B, validates inside Docker sandbox. Zero human intervention.
Enterprise RAG · Agentic
Intelligent Research Assistant
Production-grade agentic RAG pipeline with FAISS vector search, self-correcting LangGraph agent with document-relevance grading and Tavily web-search fallback.
Full-Stack · E-Commerce
MyDressUpProject
Full-stack MVC application with secure authentication, product catalogue management, and end-to-end order processing with RESTful API integration.
NLP · Domain-Specific AI
Bangla-ChatBot
Conversational AI for Bengali within the banking domain. Demonstrates multilingual NLP capability for underserved languages.
Beyond the Screen
Cultural Life
Rabindra Sangit — Performance
Performer/Dancer · Cultural Festival
Performed Rabindra Sangit at Heritage Institute's annual cultural festivals. These timeless compositions — where poetry, melody, and emotion merge — shaped my sense of depth and precision. The same quality that makes a raga beautiful is what makes an AI agent elegant.
Rangoli Competition
Participant · Office Festive Events
Participated in the office Rangoli competition at SMS Group — a meditative art form where geometry, colour, and intuition converge. A reminder that the best interfaces, like the best rangolis, are crafted with intention.
Office Cricket League
Player · Team Building
Active participant in the office cricket matches at SMS Group. Reading a spinner and debugging a LangGraph state graph have more in common than you'd expect — both require patience, pattern recognition, and calm under pressure.
The Poet
Words · Rhythm · Feeling
When the code is closed, I write poems. Not for anyone in particular — just to make sense of things. I find that the same instinct that helps me debug a pipeline — noticing what's missing, finding the right word for a broken thing — is exactly what writing poetry trains. Both are about saying something precise with very little.
Get In Touch
Let's build
something remarkable.
Open to AI engineering roles, LLM/RAG systems, agentic product teams, and conversations about the future of autonomous software. Based in Kolkata. Ready to relocate or work remotely.