ENGINEERING TEAM LEAD · AMSTERDAM

Vishwanath
Dugani.

I’m a hands-on engineering team lead based in Amsterdam. I build Python backends, AI systems, and cloud platforms.

At Simvia, I lead five engineers, set technical direction, and write production code. I work with the team and stakeholders from the initial problem through deployment and monitoring.

Explore the projects ↓Download CV ↓

02 / SOME OF MY WORK

Projects &
technical decisions.

Explore the work, my contribution, and the engineering tradeoffs. Each project includes an interactive example you can change.

SIMVIAAI / DOCUMENT INTELLIGENCE

Making sense of messy documents

A Python and LLM pipeline for turning different kinds of supply-chain documents into structured data people can use.

My roleArchitecture, implementation & platform

PythonLLMsEvaluationAWS

A few more detailsClose the notes+

The challenge

Supply-chain documents arrive in many formats. Manual review made turning them into consistent, usable product data a recurring operational burden.

What I built

I designed and built the pipeline from ingestion and parsing through classification, extraction, and summarisation. A configuration-driven layer lets non-engineers add document types without changing code.

The engineering decisions

Evaluation, guardrails, and human feedback loops are part of the system. That keeps quality visible as the variety of documents grows, while reusable components keep each new type from becoming a separate implementation.

The outcome

Structured data across roughly 500 document types, replacing manual review and making document support easier to extend.

Ingest→Classify→Extract→Evaluate

SIMVIAAI / KNOWLEDGE RETRIEVAL

Finding answers, with the source attached

Helping compliance analysts find what they need in a large collection of documents, with a clear route back to the evidence.

My roleRetrieval design & implementation

RAGEmbeddingsVector search

A few more detailsClose the notes+

The challenge

Answering supply-chain compliance questions meant manually searching a large collection of unstructured documents. Finding the evidence could take analysts days.

What I built

I built retrieval across the document estate using chunking, embeddings, and vector search, then connected it to source-grounded answers. The supporting source is shown alongside the response.

Why the source matters

For compliance work, an answer needs a path back to its evidence. Source visibility helps analysts inspect the underlying material and prepare an audit response.

The outcome

Questions that required days of manual searching could be answered in seconds, enabling same-day audit responses.

Question→Retrieve→Answer + sources

EVBOXBACKEND / CONNECTED INFRASTRUCTURE

The systems behind EV charging

Backend services, telemetry pipelines, and tools that help people operate and understand a fleet of connected chargers.

My roleTeam leadership, backend & infrastructure

PythonAsyncIOREST APIsDocker

A few more detailsClose the notes+

The challenge

EV chargers generate high volumes of telemetry. Operators need dependable device status and remote operations, and downtime can leave a driver unable to charge.

What I built

I built and scaled asynchronous, event-driven services to ingest and process charger data, alongside backend APIs for fleet status, diagnostics, remote operations, and reporting.

My role as a lead

I managed and mentored engineers while contributing to the backend and infrastructure. I also worked with developers and QA engineers on charger testing APIs, internal tools, and dashboards for understanding test data.

The outcome

Telemetry processing and fleet management capabilities used by engineering and operations, with reliability, concurrency, and performance central to the design.

Devices→Async pipelines→Fleet operations

SIMVIAAI / PRODUCT DELIVERY

Helping customers get started

Removing a manual configuration step from onboarding, so a new customer can get going without a long wait.

My roleWorkflow redesign & LLM integration

LLMsWorkflow designProduct thinking

A few more detailsClose the notes+

The challenge

Every new client depended on a manual configuration step. That delivery bottleneck slowed onboarding and held up the sales cycle.

What I changed

I rebuilt onboarding around an LLM to remove the configuration bottleneck. The onboarding work also included SharePoint integration to automate document ingestion. The goal was straightforward: help customers get started sooner.

The leadership connection

This work sits within my partnership with the C-suite on GenAI strategy: translating business goals into shipped capability and choosing applications where AI creates practical value.

The outcome

Onboarding moved from weeks to days. Removing the delivery bottleneck also shortened the sales cycle from months to days.

New customer→LLM onboarding→Ready to use

SIMVIAPLATFORM / AGENT INTEGRATION

Connecting agents to useful tools

Client-facing APIs and an MCP server, with permissions and an audit trail built into the integration.

My roleAPI design, permissions & implementation

MCPREST APIsAccess control

A few more detailsClose the notes+

The challenge

Customers need their AI agents to work with platform data and tools. Giving an agent that capability also means defining what it can access and making its actions traceable.

What I built

I built the MCP server and REST APIs exposed to clients, enabling their agents to retrieve data and invoke tools through scoped permissions and auditable actions.

The engineering principle

The integration boundary is part of the product. Access control and auditability belong in the design of an agent-facing interface.

The outcome

Client-facing integration capabilities that connect customer agents to the platform while keeping access scoped and actions auditable.

Client agent→MCP / API→Scoped tools

EFFECT PHOTONICS + PHILIPSFULL-STACK / AUTOMATION

Software for the lab and production line

Working with scientists and operators to turn experiments, measurements, and manufacturing steps into repeatable workflows.

My roleFull-stack software & research collaboration

PythonFull-stackTest automation

A few more detailsClose the notes+

The challenge

Precision hardware needs repeatable measurement and dependable workflows. Scientists need room to explore; manufacturing operators need consistency across stations.

At EFFECT Photonics

I designed and built a full-stack system for the production of SFP optical transceiver modules: executing test sequences, capturing measurements, tracking progress and yield, and coordinating station workflows. I also worked directly with physicists on device characterisation and tuning tools.

At Philips

I built Python measurement, test, and data-acquisition systems for photonics research, along with visualisation tools. Working with scientists and firmware engineers taught me to turn experimental methods into usable software.

The outcome

Manual research operations became repeatable and unattended. Production operators gained integrated tooling, and automated testing and CI/CD helped standardise builds and releases at EFFECT Photonics.

Experiment→Repeatable tooling→Production workflow

These are summaries of my contributions. There’s no proprietary code or customer data here — just a little context about the work.

03 / WORKING TOGETHER

How I work
with a team.

I lead a team of five engineers, and I still write production code. For me, those two parts of the job belong together.

Make the direction clear.

I work with the team and business stakeholders to understand the problem, talk through tradeoffs, and agree on a useful next step.

Make room for people to grow.

Design discussions, code reviews, and coaching are chances to share context and build confidence. I want teammates to have real ownership of their work.

Stay with it after it ships.

Tests, deployment, monitoring, and feedback are part of the work. I like understanding the whole system and leaving it a little easier to maintain.

TOOLS I COME BACK TO

Python / Django / FastAPI / PostgreSQL / AWS / Docker

Also: LLMs, RAG, MCP, AsyncIO, Redis, Pytest, CI/CD, Sentry, TypeScript, and React. The choice depends on the problem.

04 / THE JOURNEY

Experience.

I started with electronics and embedded systems. That took me into research, manufacturing, EV charging, and eventually the software I work on today.

2024 — NOW

Simvia

Amsterdam

Team Lead · Apr 2024 – present

Leading five engineers and staying hands-on with Python, AI document processing, compliance search, and the cloud platform. I work with the C-suite on where AI is useful, and with the team on how to build it well.

The work also includes document review tools, client-facing APIs, and SharePoint-integrated onboarding.

2021 — 2024

EVBox

Amsterdam

Senior Software Engineer & Team Lead

Building backend services and tools for connected EV chargers, while managing and mentoring engineers. A mix of telemetry, fleet operations, testing, APIs, and helping people make sense of the data.

2018 — 2021

EFFECT Photonics

Eindhoven

Senior Software Engineer

Working with physicists and production teams on software for optical transceiver manufacturing, device tuning, and research. I built full-stack tools, dashboards, and automated testing and delivery workflows.

2016 — 2018

Philips

Eindhoven

Software Developer

Building Python measurement systems, testing tools, and interfaces for research data. This was a chance to work closely with scientists and firmware engineers, and make manual experiments repeatable.

And a little further back…+

2016 / SCANIA

A thesis in high-performance computing.

At Scania, I studied Linux cluster performance and benchmarked finite element analysis and computational fluid dynamics applications for my master’s thesis.

2015–2016 / WATTY

Energy data and embedded hardware.

As an intern and part-time data integrity analyst, I worked on GSM firmware, hardware testing, and automating data cleaning for disaggregation algorithms.

2012–2014 / BRADY CORPORATION

Learning the business side.

A business analyst role working on reporting, forecasting, and financial analysis — an earlier part of my path before the move to Sweden.

2011–2012 / LI2-INNOVATIONS

Starting with a balancing robot.

As a junior developer, I worked on a two-wheeled self-balancing robot, using motion sensors, filtering, and closed-loop PID control.

EDUCATION

M.Sc. Embedded Systems

KTH Royal Institute of Technology

Stockholm, Sweden · 2014–2016

B.E. Electrical & Electronics Engineering

M.S. Ramaiah Institute of Technology

Bangalore, India · 2008–2012

05 / OUTSIDE WORK

Outside work.

I enjoy bouldering and photography outside work.

Bouldering

Photography

Open Water DiverCertification

06 / SAY HELLO

Get in touch.

I’m open to engineering team lead opportunities. If you’re building a team, working on an interesting problem, or think we might enjoy working together, I’d be happy to talk.

vishwanath.dugani@gmail.com

Find me on LinkedIn Download my CV