About Me
I'm a Firmware Engineer at Temple, building firmware and edge-AI for a forehead-worn health wearable — sensor-side signal processing, on-chip ML, and power management on the nRF52840 (Zephyr RTOS).
Previously at Cadence Design Systems, I worked on the NeoNPU AI compiler, developing compiler passes, quantization algorithms, and hardware-runnable micro-ops for edge AI acceleration. Before that, at Samsung R&D, I specialized in Android low-level performance optimization, working with the Linux kernel, memory management (cgroup v2, memcg), and kernel-space page compression mechanisms.
I'm an IIT Ropar graduate (B.Tech Electrical Engineering, 8.08 CGPA) with a strong foundation in algorithms, competitive programming (Codeforces Specialist - 1567 rating), and a passion for building systems that push performance boundaries.
Location
Noida, India
Education
B.Tech Electrical Engineering, IIT Ropar
Current Role
Firmware Engineer at Temple
Skills & Technologies
Languages
AI & ML
Systems & Tools
Embedded & Firmware
Specializations
Experience
Firmware Engineer - Firmware & Edge AI
- ›Building firmware for a forehead-worn health wearable on the nRF52840 (Zephyr RTOS), integrating the LSM6DSOX IMU and MAXM86161 PPG sensors
- ›Achieved 3333Hz sustained IMU streaming over BLE with zero data loss using only +1.5KB of additional RAM
- ›Offloaded tap, blink, and chewing detection onto the LSM6DSOX's on-chip FSM/MLC, keeping the MCU asleep until real events occur
- ›Built an I2C power manager with bus suspend/resume and sequenced peripheral low-power modes to cut idle power draw
- ›Deployed an on-chip TFLM keyword-detection model (int8 quantized) using CMSIS-DSP for fixed-point signal preprocessing
Software Engineer - Neo NPU AI Team
- ›Developed key components of Cadence's NeoNPU neural network compiler, enabling deployment of ONNX, TFLite, and MLIR models on Edge AI accelerators
- ›Implemented operator fusion, layout transformations, and quantization passes (per-tensor, per-channel, symmetric/asymmetric) in the compiler IR
- ›Built multiprocessing-based testing framework, reducing end-to-end model testing time from 28 hours to 2 hours
- ›Implemented BF16 support for tanh, sigmoid, ResizeBilinear, and Reduce operators
- ›Devised kernel decomposition algorithm for 2D kernels, improving Conv operation performance by 19% and reducing memory footprint
Software Engineer - OnDevice AI Team
- ›Designed runtime memory control framework using cgroup v2 and memcg for dynamic process-level memory management in Android
- ›Optimized kernel-space page compression with improved hash-based deduplication and lock-free execution paths (filed as patent, received A2 evaluation)
- ›Built Python analysis tools for memory profiling and system diagnostics from custom daemon data
- ›Received Samsung SPOT Award, Excellence Award, and cleared SWC Advanced & Pro certifications
Software Engineering Intern - MODEM PHY Team
- ›Developed log analysis program with Excel-like tables, color-coded logs, filtering, and error plots
- ›Optimized log processing with WxPython for stress testing and efficient filtering of large log files
Research Assistant under Dr. Arun Kumar
- ›Developed algorithmic trading solution for options derivatives analysis based on expiration, VWAP, and IV, achieving 21% return
- ›Applied time-series analysis and optimized LSTM models, reducing error rate to 0.4%
- ›Integrated macroeconomic factors to increase prediction accuracy using ensemble methods and hyperparameter tuning
Education
Indian Institute of Technology, Ropar
Achievements
Codeforces Specialist
2024Achieved 1567 rating on Codeforces with highest rank of 552, demonstrating strong algorithmic problem-solving skills
Samsung SPOT & Excellence Award
2024Honored with Samsung SPOT Award and Excellence Award for outstanding contributions to kernel optimization and memory management work
Samsung SWC Advanced & Pro Certification
2023Cleared Samsung SWC Advanced and Professional competency tests, demonstrating mastery in data structures and algorithms
JEE Advanced AIR 4373
2020Secured All India Rank 4373 in JEE Advanced, one of India's most competitive engineering entrance examinations
JEE Mains - Top 0.4%
2020Ranked in top 0.4% among 1.2 million candidates in JEE Mains examination
Indian National Chemistry Olympiad
2020Selected among 800 students across India for the prestigious Indian National Chemistry Olympiad
Department Placement Coordinator
2023Led placement initiatives for 1350+ students at IIT Ropar, successfully negotiating with companies like Texas Instruments, NVIDIA, and BPCL