smResume
SOFTWARE ENGINEER / BENGALURU, IN OPEN TO WHAT'S NEXT

Thought through.
Built all the way.

I build developer platforms, fast backends, and ambitious software you can actually use.

Explore my work
CURRENT CHAPTERSoftware Engineer I
at Honeywell

From intern to engineer.
From an idea to a platform used by ~50 teams.

PLATFORMS / BACKENDS / REALTIME SYSTEMSPORTFOLIO / 2026

Built in the real world.

HoneywellOneSol AI LabsElogix Software& upstream

01 / SELECTED WORK

Different problems.
Same intent.

Follow the problem past the interface.
Into the storage, the protocol, the failure case.
Then bring the whole thing together.

Argus / 01

A search engine and a browser engine. Built from first principles.

Java 17Block-Max WANDBM25 / nDCGWrite-ahead log

BUILT AT HONEYWELL

A tool becomes a platform
when people make it theirs.

I conceived and built AI DevKit: a VS Code and GitHub Copilot multi-agent platform connecting requirements, repository standards, specialist agents and reviewed pull requests.

~50 teams

use and customize the platform
for feature and defect work

~40%

less provisioning turnaround
through workflow automation

Relative time Before = 100
Before100
After~60
Approximate, resume-reported result.

Architecture to adoption.

Initial architecture, Python connectors,
log analysis, telemetry and interfaces.

Internal work. Public summary only.The full context

02 / THE PROJECT INDEX

Curiosity,
with a commit history.

03 / WHERE I'VE BUILT

A little further
every chapter.

Enterprise platforms. AI infrastructure. Full-stack delivery. A progression toward owning more of the system, not just the ticket.

Jul 2026 / Present · Intern Jan–Jun 2026 / Bengaluru, India

Software Engineer I

Honeywell

  • Conceived and built AI DevKit, a VS Code/GitHub Copilot multi-agent platform with MCP integrations, custom Python connectors, and repository-specific standards. About 50 teams use and customize it for feature and defect work.
  • Built AI LogKit to redact logs, correlate Rancher, Kibana and AKS evidence with source, and produce a reviewed fix pull request or a cited root-cause analysis and runbook.
  • Owned the initial architecture and implementation, with internal co-inventor recognition through Honeywell's patent process. This is recognition, not a claim of a granted patent.
  • Automated DevKit and AI LogKit provisioning, reducing turnaround by approximately 40%; added telemetry and React incident-analysis interfaces.
PythonMCPC#/.NETReactKubernetes / AKSELK / Kibana

05 / THE PERSON BEHIND THE WORK

I like the part
where it gets difficult.

Sayantan M.

I'm a software engineer in Bengaluru. I graduated in Computer Science from VIT-AP, worked with two AI-focused teams, and converted from intern to Software Engineer I at Honeywell.

My work sits between developer experience and the machinery underneath: orchestration, retrieval, APIs, storage, real-time state. I like knowing why a system works, what happens when it doesn't, and whether the next engineer can make sense of it.

Outside the day job, I build things that stretch that understanding. Sometimes it's a search engine. Sometimes it's a racing game.

9.03/10B.TECH CSE / VIT-AP / 2026
869LEETCODE SOLVED / 145 HARD

Coding profile counts from the September 2026 resume. Smart India Hackathon 2024 finalist.

PLATFORMS & AI

Python / FastAPI / MCP / Agent orchestration / Retrieval

BACKEND & DATA

PostgreSQL / Redis / Milvus / FAISS / C# & .NET / SQL

BUILD & SHIP

TypeScript / React / Next.js / Docker / Kubernetes / CI/CD

06 / A GOOD NEXT CHAPTER

OPEN TO ENGINEERING OPPORTUNITIES

LET'S BUILD
SOMETHING
THAT MATTERS.

Backend, platform, or a problem that needs both.
I'd like to hear what you're working on.

PROJECT FILE /
THE 60-SECOND BRIEF

Sayantan Mandal.

Software Engineer I / Honeywell / Bengaluru

Experience metrics are from the supplied resume. Internal platform details remain confidential; recognition is not presented as a granted patent.