Baibhav Pathak

Software Engineer • IoT • AI/ML • Cloud

I build dependable systems
that ship, scale, and stay up.

Software engineer with a taste for data-driven products. From simulated IoT fleets to real-time embedded applications, modern web dashboards, and cloud-native services — I turn raw signals into reliable, scalable software. Skilled across C++/SystemC, Python, and full-stack development with React • Node • SQL • AWS, I focus on clean architecture and measurable impact.

Experience

2+ years
Software Engineer
Current FocusAI Fleet Health Monitoring (in progress)
CloudAWS • CI/CD • Containers
Full-StackReact • Node • SQL
Baibhav profile photo
▍
See My Work Let’s Collaborate

Skills

Tools I actually use

Experience

What I shipped

Software Engineer — Packaging & Automation

ThermoFisher Scientific • Bangalore, India • Jan 2021 — Aug 2023
  • Developed & deployed 50+ software packages for macOS & Windows (AdminStudio, Shell, SCCM) with 100% seamless integration.
  • Automated test & deployment with PowerShell, Python, and VBScript, cutting manual work by 30% and improving reliability pre-UAT.
  • Resolved complex 3rd-level packaging issues with a 95% success rate, stabilizing operations for 200+ Autopilot & Azure users.
  • Built data extraction tools against Microsoft Graph / Azure DB, improving data accuracy by 35% and enabling data-driven decisions.

Education

Built on fundamentals

Northeastern University

M.S. in Internet of Things • Boston, MA, USA • Aug 2023 — May 2025
Core Subjects
  • Data Structures & Algorithms
  • Wireless Sensors and Networks in IoT
  • Mobile and Wireless Networks
  • Machine Learning & Pattern Recognition
  • Hardware/Software Co-Design
  • Network Security
View Degree

Visvesvaraya Technological University

B.Tech in Computer Science • Karnataka, India • Aug 2016 — Jun 2020
Core Subjects
  • Operating Systems
  • Machine Learning
  • Computer Networks
  • Python
  • Computer Architecture
View Degree

Projects

Selected Work

AutoPulse — AI Fleet Health Management

in progress

Simulated OBD-II telemetry + LSTM forecasts to flag anomalies before breakdowns. Focus on downtime reduction via targeted alerts and failure signatures.

  • Streamed sensor frames over MQTT; feature windowing for sequence models.
  • LSTM baseline; exploring autoencoder thresholds for anomaly scoring.
  • Operational dashboard (React) with trends, alerts, and drill-downs.
PythonLSTMMQTTReact PostgreSQLAWS
View Code

Vision Transformer SoC

SystemC SoC for image classification with modular patch embeddings, transformer blocks, and MLP head; explored fixed-point quantization for edge.

  • Integrated pretrained ViT weights & parallelized embedding pipeline.
  • Fixed-point arithmetic sweep for FPGA feasibility.
  • Clock-accurate simulation & trace for throughput/latency study.
SystemCC++Vision Transformer Fixed-PointFPGA (feasibility)
View Code

ASL Interpretation Tool

Real-time ASL interpreter using OpenCV + MediaPipe with >90% accuracy; boosted via Random Forest classifier and lightweight web UI.

  • Key-point extraction pipeline with temporal smoothing.
  • Model comparison (SVM vs RF) — RF improved accuracy ~20%.
  • Browser preview for immediate feedback.
TensorFlowOpenCVMediaPipe Random ForestPythonHTML/CSS/JS
View Code

Boston Subway Network Visualization

Interactive Processing sketch for MBTA routes with shortest-path planning and live data refresh from Python scraper.

  • Graph model with stations/edges; Dijkstra for path queries.
  • CSV ingest from MBTA feeds; basic delay annotations.
  • Zoom/pan interactions & color-coded lines.
ProcessingPythonGraph Theory Shortest PathMBTA
View Code

Cellular Network Drone

Quadcopter with extended range via cellular towers; optimized SPI/I²C transfer (~25% faster) and onboard camera streaming.

  • Arduino flight control + Python telemetry bridge.
  • Network stack tuning for uplink stability.
  • Modular sensor bay for field experiments.
ArduinoPythonC++ SPI/I²CNetworkingCamera

Personal Portfolio (this site)

Dark, glassy, responsive portfolio with theme toggle, typing line, particles, and snap-scroll sections. Built for clarity and speed.

  • Semantic HTML + modern CSS + vanilla JS.
  • Keyboard-friendly nav and reduced-motion fallbacks.
  • Modular sections for easy iteration.
HTMLCSSJavaScript AccessibilityResponsive
Source Code

Achievements

Results that matter

Contact

Let’s build something
GitHub @pathakbaibhav
LinkedIn /in/baibhavkp
Instagram @baibhav_pathak