Tanay Anand
Building at the intersection of ML systems, embedded firmware, and distributed infrastructure. Currently researching kernel optimization for neural network workloads.
Where I've Worked
- Built Python and Java tools to analyze distributed telemetry across 15+ networking services, reducing debugging time for production incidents by 40%.
- Implemented validation modules that processed 10M+ telemetry events to detect inconsistencies in system state and network metrics.
- Traced execution paths across distributed services to isolate software defects and verify intermediate computation stages.
- Developed automated test pipelines to validate telemetry processing logic and improve reliability of internal monitoring infrastructure.
- Developed a PyTorch-based notes assistant that processed 5K+ documents, improving model accuracy by 20% through data preprocessing and validation workflows.
- Built testing utilities to verify ML data pipelines, catch data handling issues, and improve reproducibility of model inference behavior.
- Investigated runtime and memory-related failures in model execution, resolving edge cases in data loading and preprocessing.
- Taught Python, Java, and C to 50+ students, with emphasis on software design, algorithmic thinking, and systematic debugging.
- TA for the undergraduate Embedded Control course covering microcontroller programming, interrupt-driven I/O, real-time scheduling, and hardware-software co-design.
- Held office hours and lab sessions guiding 60+ students through hands-on firmware development on ARM-based platforms using C, GPIO, timers, and finite state machines.
- Designed and graded lab assignments involving sensor integration, motor control, and real-time event handling on resource-constrained hardware.
- Debugged student hardware setups and firmware, troubleshooting issues across serial communication, interrupt priority conflicts, and timing-sensitive control loops.
Academic Background
Rensselaer Polytechnic Institute
Machine Learning • Operating Systems • Distributed Systems • Computer Networks • Data Structures & Algorithms • Parallel Computing • Embedded Control • Computer Architecture
Carnegie Mellon Pre-College
Programmer (Unity C#), Developer & Audio Engineer. Built game prototypes in collaborative team environments using industry-standard tools.
Tech Stack
Things I've Built
Mini-Matching-Engine
A simplified electronic trading engine written in pure C for Linux. It accepts buy and sell limit orders from stdin or a file, maintains an in-memory order book, and matches orders using price-time priority
Transformer Scaling Explainer
Interactive tool modeling Transformer inference scaling. Adjust layers, sequence length, precision, and hardware to visualize latency, attention cost, and KV cache behavior in real time.
Neural Networks & LLM Visualizer
Explore tokenization, attention patterns, and internal representations of large language models with real-time visual feedback.
ASIC AI Agent - Claude
Built an AI agent for RTL verification that reads Verilog, generates testbenches, runs simulations, parses results, and reports pass/fail automatically.
Kernel Optimization Study
Benchmarked matmul and convolution kernels across CPU/GPU. Achieved 35% training speedup via memory access optimization and tensor-level tuning.
Robotics Control System
Embedded C firmware with FSM control logic, timers, interrupts, GPIO, and event-driven state management for real-time sensor-driven behavior.
Telemetry Validation Toolkit
Python toolkit to parse, validate, and analyze structured system logs. Modular architecture for reusable validation logic across datasets.
Distributed Key-Value Store
Raft consensus-based KV store with linearizable reads/writes, leader election, log replication, and snapshotting for fault tolerance.
Packet Sniffer & Analyzer
Low-level packet capture decoding Ethernet, IP, TCP/UDP headers with real-time traffic visualization and filtering by protocol, port, and address.
Custom Memory Allocator
User-space allocator with first-fit, best-fit, and buddy strategies. Benchmarked fragmentation and throughput against glibc malloc.