mc.Maneet Chatterjee
CV

Mechanical engineering · Intelligent systems

Maneet
Chatterjee.

Building intelligent systems
that work in the physical world.

I work at the intersection of robotics, AI, and aerospace engineering — connecting learning-based methods with simulation, physical constraints, and rigorous validation.

Explore my researchView CV
Research & R&D experience
SynopsysUniversity of TsukubaAnsysIIEST Shibpur

01 / Selected work

From ideas to validated systems.

Generative engineering, embodied intelligence, and perception — with an emphasis on how models behave beyond training.

Generative design · Aerospace01

Physics-gated generative CAD
for aircraft inverse design

A Python/SUAVE pipeline coupling probabilistic design grammars and Sobol sampling with mission analysis, conservation-law checks, and automated 3D geometry.

30physical constraints
21/21valid, watertight CAD designs
in the evaluated set
SUAVEPythonGenerative CADPhysics validation
Method & validation

Combines component-aware design generation with converged mission analysis and watertight STL/OBJ export. The evaluated set includes 11/11 validated turbofan solutions; climb fuel was optimized to 10.2% of maximum takeoff weight.

Selected project · September 2026. Results describe the evaluated designs.

Robotics · Embodied AI02

Vision-language-action
robotic manipulation

A CLIP-guided pick-and-place pipeline in PyBullet, connecting language-conditioned object perception to manipulation and confidence-aware evaluation.

5 / 4objects / sorting zones
0.75ROC-AUC for execution
outcome separation
CLIP ViT-B/32PyBulletPython
Method & evaluation

Uses the confidence margin between target and distractor objects to quantify semantic ambiguity. Successful and failed executions differed under a Mann–Whitney test (p < 0.01).

Simulation study. ROC-AUC measures outcome separation.

Mechanical design · Reinforcement learning03

MoonBot:
rover design to learned locomotion

A six-wheel rover with an articulated arm, developed from SolidWorks CAD through URDF modeling to a Gym-compatible reinforcement learning environment.

SolidWorksPyBulletPPOGymnasium
Design & training

Models inertial properties, joint constraints, and collision geometry for multi-contact dynamics. PPO locomotion uses mass and friction randomization to address simulation uncertainty.

Simulation-based development; physical robot deployment remains future work.

Explore repository
Multiphysics · Thermal engineering04

Cross-solver CFD validation
of curved cooling ducts

Matched conjugate heat-transfer cases in Fluent, COMSOL, and OpenFOAM, investigating Dean-type vortices and thermal non-uniformity.

+18%mean Nusselt number
vs. project baseline
6%pressure-drop agreement
across solvers
FluentCOMSOLOpenFOAMk–ω SST
Model & application

Solves steady incompressible Navier–Stokes and energy equations with k–ω SST turbulence modeling. The study supports thermal design for EV batteries, avionics, and aerospace environmental-control systems.

Selected project · May 2026. Improvements are relative to the project baseline.

02 / Papers & preprints

Selected publications.

Google Scholar
2026
PreprintarXiv · September 2026Equal contribution

Temporally Ordered Region-Token Mamba with Logit-Space Diffusion for Remote Sensing Change Detection

Anuvab Sen, Maneet Chatterjee, Aparup Ghosh, Udayon Sen, Arnav Aditya, Yixin Zhang

BMD-CD combines temporally ordered region tokens, state-space propagation, feature disentanglement, and five-step logit-space diffusion for precise change localization.

PaperPDFCode
2026
CVPR workshopMONTI · Non-archivalFirst author

Topology-Constrained Graph-Mamba with Logit Diffusion for Change Detection

M. Chatterjee, A. Ghosh, A. Sen, U. Sen

PaperWorkshop
2025
IEEE InGARSSConference paper

Beyond CNNs: EfficientMamba Powered Sequence Learning and Curriculum-Guided Domain Adaptation for Fossil Classification

A. Roy, A. Sen, M. Chatterjee, S. Das, J. Paul

2025
IEEE IGARSSConference paperFirst author

EEMAMBA: A Hardware-Aware Energy-Efficient State-Space Model for EuroSAT Classification

Maneet Chatterjee, Anuvab Sen, Subhabrata Roy

Paper / DOICode
2024
IEEE CSNTConference paperFirst author

ExoSpikeNet: A Light Curve Analysis Based Spiking Neural Network for Exoplanet Detection

Maneet Chatterjee, Anuvab Sen, Subhabrata Roy

Paper / DOICode

03 / Research & industry

Engineering across disciplines.

Jun 2026 — PresentPune, India

Synopsys R&D Intern

Developing generative inverse-design methods for conceptual eVTOLs, with a canonical aircraft schema encoding components, configurations, topology, and placement rules.

Coupled conditional latent diffusion to SUAVE mission analysis, native OpenVSP geometry, and VSPAERO aerodynamics. 91% of generated designs passed geometric and forward-model checks, with solver convergence, validity, and provenance logged.

Inverse designLatent diffusionOpenVSP / VSPAERO
Certificate
Mar 2026 — PresentTsukuba, Japan

University of Tsukuba Research Assistant

Building MedSAM/SAM 3 and VideoMAE pipelines to localize surgical stapler events across 28 laparoscopic procedures, 142 annotated events, and 14,776 frames.

Fused frame features, clip representations, region motion, and hard negatives; achieved 91.6% balanced accuracy and 89.7% F1 under five-fold case-disjoint cross-validation, with calibration and safety-rubric assessment.

Surgical video understandingSegmentationModel validation
May — Jul 2025Pune, India

Ansys R&D Intern

Developed vision-guided robotic manipulation using OpenCV, PyTorch, and ArUco: object detection, camera calibration, coordinate transforms, and robot kinematics.

Integrated embedded controllers with MQTT/WebSocket monitoring on Raspberry Pi and validated behavior in Isaac Sim, Blender, and Unreal Engine.

RoboticsDigital twinsEmbedded systems
Certificate
Aug — Dec 2024Kolkata, India

IIEST Shibpur Research Assistant

Modeled solar-powered hydrogen production from measured irradiance and weather data, optimizing photovoltaic–electrolyzer power matching and system performance.

Energy systemsPhysical modeling
Research report

04 / Background

Education & recognition.

Aug 2023 — May 2027 · Expected

Indian Institute of Engineering Science and Technology Shibpur

B.Tech. in Mechanical Engineering

3.63/4.00CGPA through Semester VI

Honors & presentations

  • 2025
    GAABESU Research Award
  • 2025
    IEEE GRSS Travel Grant

    For IGARSS

  • 2025
    Ansys TechCon & AIIoT

    Co-authored / presented work on IT–OT integration and digital twins at the 6th Ansys TechCon and 2nd AIIoT.

Technical toolkit

Learning & computation

Python · C++ · MATLAB · SQL
PyTorch · TensorFlow · OpenCV
scikit-learn · Git · Docker

Robotics & simulation

ROS 1/2 · NVIDIA Isaac Sim
PyBullet · MuJoCo · Gazebo
Simulink · Raspberry Pi · MQTT

Design & multiphysics

SolidWorks · CATIA · Fusion 360
Ansys Mechanical · Abaqus
Fluent · OpenFOAM · COMSOL

Aerospace & digital engineering

SUAVE · OpenVSP · VSPAERO
Digital twins · IT–OT integration
InfluxDB · Grafana · LaTeX

Selected certifications

Deep Learning Specialization 2026 / SOLIDWORKS Professional 2025 / MATLAB Deep Learning 2024

05 / Let's connect

Better systems start with
good conversations.

I’m interested in graduate research, research collaborations, and industry R&D opportunities in robotics, AI, and autonomous engineering.

maneet2018@gmail.com
View my CV Google Scholar GitHub LinkedIn