Sakshi B

Sakshi Bhatia

MS Robotics @ Northeastern University
Aspiring to build generalizable robots.

Profile Summary

Robotics researcher specializing in robotic manipulation, working at at the HelpingHands Lab. I am passionate about applying novel visual methods, and generative policy algorithms to solve the complex challenges of operating robots in uncertain, unseen environments.

Technical Arsenal

Languages: Python, C/C++, MATLAB, SQL
Frameworks: PyTorch, TensorFlow, Isaac Sim, Robosuite
Sim/Tools: Mujoco, CARLA, Fusion360, Arduino
Libraries: OpenCV, Gymnasium

Perception Suite

Augmented Reality C++ pipeline

March 2026

Augemented Reality on custom image targets with shaded 3D overlays.

  • Intrinsics & Extrinsics: Calibrated camera metrics and estimated real-time checkerboard pose for board tracking.
  • Virtual 3D objects: Created Phong-shaded 3D models (Eiffel Tower, UFO, etc.) using GLFW/OpenGL with a runtime GUI selector, virtual objects rendered onto the board.
  • Robust Feature Tracking: Extended AR to arbitrary flat targets using ORB feature matching and RANSAC homography, removing checkerboard dependency. View code

Deep Networks for handwritten character recognition

April 2026

Building, training, and analyzing deep convolutional neural networks and Vision Transformers for image recognition.

  • Model Training & Evaluation: Developed and evaluated a CNN baseline and a Vision Transformer on MNIST, achieving 99.22% classification accuracy.
  • Transfer Learning on Characters: Adapted the pre-trained backbone to classify custom characters, optimized the pipeline to overcome real-world domain gaps.
  • Systematic Analysis & Optimization: Conducted a 100-configuration hyperparameter sweep across 10 dimensions, and visualized learned features using pre-trained ResNet-18 layers and a fixed Gabor filter bank.

Experience

Graduate Research Assistant | HelpingHands Lab, NEU

Feb 2025 – Present | Boston, MA
  • Built a GELLO arm to record demonstration data; reproduced Diffusion Policy in simulation and on UR5 arm.
  • [In Progress] modified Diffusion Policy combining VLMs for visual generalization
  • Focus: Robotic Manipulation, Sample Efficiency, Generalization.

Teaching Assistant | EECE Dept, NEU

Sep 2025 – Dec 2025 | Boston, MA
  • Assisting for EECE5550 Mobile Robotics (Perception, Planning, Control).
  • Assembled TurtleBot3 Burger and validated hardware/software stacks.

Product Analyst | Sprinklr

July 2022 – Jan 2023 | Bangalore, India
  • Deployed regex-based automation for content filtering on child-safe platforms, reducing manual work by 93%.
  • Managed 15+ global clients and executed 40+ data insight dashboards.

Research Intern | NIT Jalandhar

July 2021 – Sep 2021 | India
  • Developed a Python package for 'Artificial Electric Field', a novel optimization algorithm published on PyPI.
  • Visualized convergence for multi-dimensional non-linear problems.

Projects

Diffusion-guided Intrinsic Curiosity

Oct 2025 – Nov 2025

Solving long-horizon tasks with minimal demonstrations.

  • Implemented Intrinsic Curiosity Module (ICM) for sparse environments.
  • Trained a diffusion model to learn state sequences while ICM predicts actions.

Bipedal Walker (PPO + Attention)

Feb 2025 – April 2025

Locomotion stability for uneven terrain traversal.

  • Solved OpenAI Gym’s Bipedal-Walker (Avg Reward: 317) without falling.
  • Implemented PPO with attention/cross-attention mechanisms and performed ablation studies.

Education

Northeastern University

Sept 2024 – Dec 2026 | Boston, MA

MS in Robotics

GPA: 3.88/4.0

Focus: Advanced Computer Vision, Reinforcement Learning, Geometrical Deep Learning.

National Institute of Technology, Jalandhar

Aug 2018 – May 2022 | India

Bachelor of Technology

GPA: 8.09/10

Focus: Soft Computing, Mechatronics, Data Structures

Accomplishments & Activities

Achievements

  • Best Android XR App with Gemini Live API by Google DeepMind. Reality Hack 2026, MIT Media Labs
  • All India Rank 1243 / 39210 (Percentile: 96.83). GATE Data Science & AI (2024)
  • Cracked Post Baccalaureate Fellowship Exam (IIT Madras, Oct 2023). Robert Bosch Centre for Data Science and AI
  • Best Intern Award: Hybrid Electric Vehicles Internship (Dec 2018).

Published Research

Fuzzy-based techniques for clustering in wireless sensor networks (WSNs): Recent advances, challenges, and future directions.
DOI (Wiley)

Auxiliary Missions

  • Transmission & Fabrication Engineer (Team Avishkar, 2018-2019): Designed gear characteristics, simulated transmission in MATLAB-Simulink, fabricated vehicle body using welding/cutting techniques.
  • Google Developer’s Student Club (2020-2022): Organized workshops and mentoring sessions in Machine Learning.
  • Entrepreneurship Cell (2019-2022): Official social media manager; actively created awareness on entrepreneurship.