Sanskar Jadhav

About

I specialize in deploying learning pipelines for real-world applications of human-assisted tasks, such as adaptive navigation, visual understanding, and system control. My work and interest lies in embodied cognitive AI, a.k.a. "teaching robots to be more human in nature".

Skills

• Technical: Python - PyTorch, TensorFlow, OpenCV, ROS 2, MilvusDB, SQL, Neo4j, MERN Stack, Docker, Git, Linux, MATLAB, C/C++, Java
• Professional: Isaac Sim, NVIDIA Remembr, Unitree SDK, Meta Segment Anything 2, Voilà Dashboards, Streamlit, Ollama, ROS 2 Services, Model Optimization and Pipeline Automation

Education

Cornell University - Cornell Tech

MEng in Electrical & Computer Engineering

2025 – 2026

Merit Scholar.

Symbiosis International University - Symbiosis Institute of Technology

BTech in Artificial Intelligence & Machine Learning (Minor in Aerial Robotics) - CGPA 9.45 / 10

2021 – 2025

Graduated Summa Cum Laude (Rank 1).
Individual final year project on autonomous AI for robot navigation and visualization.

Vibgyor High

Cambridge International AS & A Levels - 5 A*s

2019 – 2021

Graduated Valedictorian.

Experience

AI/ML Intern – Bharat Forge Limited, Pune, India

Aug 2024 – Jul 2025 (One-year internship, offered FT Engineer role, opted for pursuing Masters)

• Worked in the Robotics team at the AI & ML Tech Centre on developing core AGI modules
• Developed a custom local application utilizing Meta’s SAM 2 model and YOLO for autonomous segmentation and annotation of multiple objects in a video, eliminating the need for online annotation tools and maintaining data privacy
• Created a docker for real-time angle detection with streaming using ROSBridge for anomaly detection in machinery
• Created a robust open area exploration algorithm using vision with dynamic obstacle avoidance tested on Unitree quadruped, wheeled, and humanoid robots both indoors and on shop floors
• Developed a custom pipeline integrating NVIDIA’s Remembr with deployment on the Unitree B2 robot using OpenAI for environment visualization, MilvusDB for memory embeddings, and Llama3 for similarity matching when querying
• Worked on inter-robot communication and Sim2Real imitation using MQTT and ROS topics in Isaac Sim with the Unitree Go2 and UR10 arm robots

Projects

Multimodal Fusion for AQI prediction using UAV onboard sensor data with CNN-LSTMs

Jul - Nov 2024, Robotics & Automation Dept., SIT

Optimization of Q-learning Networks for Efficient UAV Swarm Surveillance with Dynamic Resolution Constraints

Jan - Apr 2024, Robotics & Automation Dept., SIT

Bird's Eye: LSTMs for Flight Trajectory Prediction and Airport Weather Forecasting for Advancing Autonomy in ATC Operations

Feb - Apr 2023, AI & ML Dept., SIT

Awards & Honors

Certificate of Excellence - Bharat Forge

May 2025

Handed to me by Group CIOs Mr. Abhijit Moharir and Mr. Yogesh Zope for my simulation setup in Isaac Sim.

Judge at InterIIT Tech Meet 13.0 - IIT Bombay

Dec 2024

Designed the problem statement of "Centralized Intelligence for Dynamic Swarm Navigation" from Bharat Forge and evaluated participating undergraduates from all 23 IITs.

Certificate of Appreciation - Bharat Forge

Sep 2024

Handed to me by Group CIOs Mr. Abhijit Moharir and Mr. Yogesh Zope for my performance in the Robotics team.