$
M S Gulshan — Senior AI Software Engineer
I build end-to-end AI solutions — fine-tuned LLMs, production-grade pipelines and automation — and lead the team and client delivery behind them.
View my work01 — About
About me
I'm a Senior AI Software Engineer at Vrtta Green Solutions, where I design and ship end-to-end AI solutions: fine-tuning task-specific LLMs with LoRA/QLoRA, building industry-grade pipelines that automate manual workflows, and engineering the production software around them. I also lead the team's planning and development and manage our client engagements, from first requirements to delivery.
My latest system automates product carbon footprinting end to end: it turns a manufacturer's messy, multi-page parts lists into emission reports, with fine-tuned local LLMs parsing at 98.7% accuracy and search matching materials across a library of 1M+. On my own time I built AnswerGate, an open-source tool that fails a pull request when an AI assistant's answers get worse. Before that I was a Trainee Decision Scientist at Mu Sigma and a MERN stack intern at EY Global, and built an LSTM-based adaptive traffic signaling system with Bangalore Traffic Police.
I graduated in Information Science & Engineering from Global Academy of Technology in 2025 as the batch Gold Medalist (CGPA 9.73).
- 98.7%
- Accuracy of fine-tuned LLMs in production
- 1M+
- Materials matched automatically
- 40%
- Lower retrieval latency
02 — Career
Work experience
Vrtta Green Solutions
Senior AI Software Engineer
Promotion path
- Lead the team's planning and development and manage client relationships, from scoping requirements to delivering production releases.
- Fine-tuned open-source LLMs with LoRA/QLoRA and deployed task-specific models for parts-list parsing and material classification.
- Architected the AI pipeline behind a product carbon footprint engine — async ingestion parses unstructured, multi-page parts lists at 98.7% accuracy, automating a manual sustainability-reporting workflow.
- Built hybrid Qdrant search (dense vectors + metadata filters) that maps each material to life-cycle-assessment libraries across 1M+ entries.
- Scored every categorisation with fuzzy matching and multi-LLM agreement, and cut retrieval latency 40% with token-budget throttling and NDJSON/SSE streaming from Flask.
Mu Sigma
Trainee Decision Scientist
- Engineered data-processing pipelines and analytical models that turned complex enterprise datasets into actionable business insights.
- Applied Mu Sigma's structured decision-science problem-solving to support enterprise-level decisions.
EY Global (via Edunet)
MERN Stack Developer Intern
- Developed scalable web applications handling 100+ concurrent users while keeping performance optimised.
03 — Selected work
Featured projects
01 / 2025 · AI/ML · Data · Full-Stack
Product Carbon Footprint Engine
Fine-tuned LLMs read a manufacturer's raw parts list, match every material against a 1M+ record emissions library and produce the product's carbon footprint report — automatically.
- 98.7%
- Parts-list parsing accuracy
- 1M+
- Materials searchable
- 40%
- Lower retrieval latency
02 / 2026 · AI/ML · Full-Stack
AnswerGate — Regression Tests for AI Assistants
An open-source CI check that catches the moment an AI assistant's answers get worse after a code change, and fails the pull request with the exact questions that got worse.
- 36/42
- Injected faults caught
- 0
- False alarms, 1,070 questions
- 19/44
- Questions flagged by one check
Watch · 1:35 · sound on
03 / 2026 · Full-Stack
Exsy — Family Money Book
An installable money app for a family: separate books per person, EMIs with GST, loans, a chit fund with fair draws, and entries filled in from a pasted bank SMS.
Watch · 1:03 · sound on
04 · AI/ML · Data
Adaptive Traffic Signaling — with Bangalore Traffic Police
LSTM-based traffic optimisation engine that learns from high-frequency, real-time sensor streams to adapt signal timing, built in collaboration with Bangalore Traffic Police.
- 30%
- Less congestion (simulated)
04 — Work with me
End-to-end software
I build complete software solutions from scratch to deployed — design, engineering, AI and launch, with one point of contact the whole way.
- 01
Discover & plan
We pin down the problem, the users and what success looks like.
- Requirements
- Architecture
- Timeline
- 02
Design
Flows, screens and system design you can review before any code.
- UX flows
- UI screens
- System design
- 03
Build
The complete product — backend, APIs, database and front end.
- Python APIs
- React front end
- Databases
- 04
Add the AI
Fine-tuned LLMs, pipelines and automation built into the product.
- Fine-tuned LLMs
- AI pipelines
- Automation
- 05
Deploy
Shipped to production with tests, CI/CD and monitoring.
- Cloud or on-prem
- CI/CD
- Monitoring
- 06
Support
Docs, a walkthrough for your team and improvements after launch.
- Documentation
- Handover
- Iteration
Free project plan
Tell me what you need. Get a plan in 48 hours.
A free one-page plan for your project, written for your idea. No commitment.
- Scope and features
- Architecture and tech stack
- Timeline and milestones
05 — Capabilities
Skills & tools
AI Engineering
LLM Fine-tuning (LoRA / QLoRA)/ AI Pipelines & Automation/ Local LLM Deployment/ RAG & Vector Search/ Prompt Engineering/ LSTM Networks
Backend
Python/ Flask/ SQL/ REST API Design/ Async Ingestion Pipelines
Frontend
React.js/ Server-Sent Events/ Streaming UIs/ UI/UX Optimisation
Data & Tooling
MongoDB/ Git/ VS Code/ OpenCode/ Agile/ SDLC
06 — Education
Education & recognition
2021 — 2025
Global Academy of Technology
B.E. in Information Science & Engineering
CGPA 9.73 · Gold Medalist
2019 — 2021
The Master's PU College
PUC in Science (PCMB)
93.6%
01
Gold Medalist
Class of 2025, B.E. Information Science & Engineering — Global Academy of Technology (CGPA 9.73)
02
National Rank 3
International Science Talent Examination (ISTE), India — 2019
07 — Contact
Get in touch
Hiring for an AI role, planning a project, or just want to talk ML — my inbox is open.
gulshanms.sde@gmail.comReplies within 24 hours