Hyderabad, India • Embedded Systems, Control & Applied Maths and Physics
Software engineer building where physics, mathematics, and low-level code intersect.
I write software that interacts directly with the physical world—from C/C++ onboard flight guidance firmware and Extended Kalman Filters to iterative ballistic solvers, multi-sensor fusion pipelines, and custom Yocto Linux builds.
By nature, I’m a fast-adapting generalist. Whether a problem calls for bare-metal Xilinx Zynq bridges, synthetic 3D camera simulation in Unreal Engine, or training, implementing, and integrating AI/ML models into production pipelines, I pick up new tools quickly and focus on getting the underlying fundamentals right.
Work Experience
Career progression across Embedded Software, Application Engineering, and Defence & Aerospace R&D.
Unistring Tech Solutions
Aerospace & Defence Systems R&D
Application Engineer
Sep 2026 – PresentLead development across multi-sensor data fusion, radar track filtering using Extended Kalman Filters, iterative ballistic solvers, and embedded C/C++ software for custom Linux-based systems. Contribute to system integration, algorithm development, and end-to-end engineering deliverables across defence and aerospace projects.
Jr Application Engineer
Sep 2023 – Sep 2026Developed autonomous drone guidance firmware in C/C++, implemented 3D coordinate transformations for ballistic integration, and built iterative drag-based ballistic solvers from mathematical first principles. Worked across embedded systems, sensor integration, and real-time engineering applications.
Embedded Software Engineering Intern
Sep 2022 – Sep 2023Selected through campus recruitment during the third year of B.Tech. Trained in PIC microcontrollers and C socket programming, then progressed to developing a Xilinx SDK Radar Target Simulator. Worked on embedded C, high-speed data handling, and memory bottleneck resolution.
Selected Engineering Work
Core R&D systems built across autonomous guidance, multi-sensor fusion, state estimation, and embedded Linux.
Autonomous Guidance & Gimbal Tracking (DKS)
Wrote the complete onboard C/C++ firmware running on a Raspberry Pi Compute Module to command an interceptor drone through autonomous takeoff, target acquisition, and Proportional Navigation (PN). Integrated a 3-axis wide-angle (83° FOV) gimbal camera with custom PID velocity control loops to maintain visual target lock regardless of airframe roll or pitch.
Multi-Sensor Fusion & Ballistic Solvers (L70 & VMCSDS)
Integrated asynchronous data streams from Radar, RF Drone Detectors (RFDD), and EO/IR PTZ cameras with automated defense launchers. Formulated 3D trigonometric transformations to eliminate parallax and leveling offsets, and built an iterative forward-computable ballistic solver accounting for aerodynamic drag to predict target lead and launch angles in real time.
Radar Track Filtering & Dual-Antenna GPS Heading
Tuned Extended Kalman Filters (EKF) using radar covariance matrices and drone kinematic limits to turn jumpy DSP track outputs into smooth, physically realistic trajectories. Also configured dual-antenna GPS baselines using moving and infinite smoothing averages to get reliable heading angles immune to magnetic interference from nearby motors.
3D PTZ Camera & Target Simulator
Built a synthetic PTZ camera and drone simulator in Unreal Engine, applying my Blender 3D modelling background for visual realism. When fed simulated radar target Lat/Lon coordinates over sockets, the virtual camera rotates to point at the target bearing, handles optical zoom, and runs closed-loop visual tracking on the 3D drone.
Custom Minimal Yocto OS & High-Speed Xilinx Zynq Bridge
Replaced bloated desktop Linux images on the Raspberry Pi Compute Module + Waveshare board by building a minimal custom OS with Yocto. Compiled my C++ guidance firmware directly into the minimal build, cutting flashing time from 35–40 minutes down to 30 seconds.
Wrote embedded C firmware on Xilinx Zynq SoCs where the Processing System (PS) acts as a high-speed bridge between FPGA RTL (over AXI) and PC DSP software (over Ethernet), ensuring zero-drop streaming of high-rate binary IQ radar data and MCU hardware control.
Workflow & Technical Versatility
Practical engineering tools, first-principles thinking, and adaptable problem-solving.
Confident with the Linux Command Line
Well versed in Linux command-line tools for debugging,
networking, system inspection, and deployment. I use utilities
such as ssh,
scp,
netcat (nc),
socat, and
cat when the task calls for
them—from inspecting socket traffic and bridging network
streams to troubleshooting embedded Linux systems and
deploying builds.
First-Principles Thinking & AI/ML Adaptability
I focus on understanding the mathematics, physics, and constraints underlying a problem before choosing an implementation. I use modern LLMs to accelerate development where useful, while reviewing and verifying the resulting code. Combined with hands-on AI/ML experience, this approach helps me quickly learn unfamiliar tools, train and implement models, and integrate them into larger software systems.
Achievements & Recognition
2x Consecutive Employee of the Year
Awarded Employee of the Year for two consecutive years in recognition of end-to-end ownership across critical defense R&D deliverables, from mathematical ballistics and sensor fusion to embedded flight software.
First Prize — Developers-Go Hackathon
Secured 1st place by rapidly architecting and shipping a complete working software prototype under tight competition deadlines.
Education
B.Tech in Computer Science & Engineering
Nalla Narasimha Reddy Group of Institutions (NNRG), Hyderabad
Intermediate Education (MPC)
Narayana Junior College, Hyderabad
Secondary School Certificate (Class X)
St. Joseph's Public School, King Koti, Hyderabad
Beyond Work
3D Modelling & Rapid Fabrication
I started 3D modelling in Blender during college and later bought my own 3D printer, expanding into mechanical CAD with Autodesk Fusion 360. Drafting 2D sketches, extruding them into 3D models, and printing functional parts at home is something I regularly do for fun and prototyping.
Speedcubing
Away from the screen, I’m practicing to become a faster speedcuber, with a current personal best of around 50 seconds.