Active Core: Offline-First & Parser Systems

I build low-latency structures.

Building low-latency compiler tooling, ultra-fast lexical parsers, and real-time offline-syncing audio architectures. I specialize in bridging the gap between low-level performance systems and responsive frontend applications. From implementing NTP-like acoustic time-sync protocols over WebSockets and Bluetooth Low Energy (BLE) to optimizing compiler lexer tokenizers for maximum throughput, I design software that is robust, fast, and completely offline-first.

View Systems Contact

// SYNTAX PARSER ROOT CHECK
[INIT] loading tokens...
TOKEN_TYPE: IDENTIFIER("PartySyncService")
TOKEN_TYPE: KEYWORD("class")
TOKEN_TYPE: BRACE_OPEN
  _serverClockOffset: 0
  _roundTripTime: 0
  _isAcousticSyncing: false

[CALIBRATING] consensus ping...
  tSelf: 17294729104
  tCross: 17294729148
  latency -> 22ms
  consensus offset calculated -> -14ms
  status: LOCKED

[COMPILING AST]
  └── Program
      └── ClassDeclaration (PartySyncService)
          ├── Field (wsChannel)
          └── Method (broadcastPlay)
              └── CallExpression (compensatedPlayback)

Featured Systems

Source files / AST outputs / DSP controllers

AudioSync (Spotify Killer)

ID: audiosync_core

A high-performance offline-first audio synchronization system. Features rapid acoustic consensus calibration using high-frequency microphone chirps to align multi-device playback down to sub-10ms offsets. Built-in BLE frame broadcast for offline party sync and local SQLite caches.

Consensus Offset< 10ms

Sync MediumBLE / WAN

DSP BuffersCircular Mono

Flutter/DartWebSocketsBluetooth BLERustSQLiteDSP (SoLoud)

Source Live Demo

Lexer & Tokenizer Engine

ID: lexer_core

A custom lexer supporting dynamic token mapping, string-to-token lexical conversions, and state-machine scanning. Built to generate abstract syntax trees (AST) with exceptional throughput, providing deep error boundary checks and native syntax highlighting structures.

Throughput1.2 GB/s

Token TypesDynamic Mapping

Parser DepthRecursive

TypeScriptRust/WASMLexing SystemsAbstract Syntax Trees (AST)State Machines

Source Live Demo

Aetheric Sound Streamer

ID: aetheric_core

Constructed with low-latency Web Audio API nodes, circular sample buffers, and fast discrete Fourier transforms for high-fidelity interactive spectrum visualization. Seamlessly handles streaming fallback states.

Buffer Latency24ms

FPS Render60fps Canvas

Sample Rate48kHz

React/Next.jsWeb Audio APIWeb WorkersCanvas APITypeScript

Source Live Demo

Execution Log

Professional history & compiler pipeline milestones

Lead Systems & Frontend Engineer

@ Acoustic Networks & Streams

2024 - Present

India

Designed and implemented high-performance real-time audio synchronization protocols using BLE packet frames and NTP-like WebSocket delay estimation.
Optimized frontend audio playbacks and DSP engines in distributed mobile/desktop structures, reaching consensus playback alignment in < 10ms.
Engineered offline caching layers utilizing SQLite and local JSON indexes to allow full application usage under extreme network constraints.

FlutterDartTypeScriptWebSocketsBLESQLite

Senior Compiler & Tooling Developer

@ LexiTech Tooling Corp

2022 - 2024

India

Built a modular, highly extensible programming language lexer and tokenizer supporting custom token types, state-machine scanning, and strict parser boundaries.
Integrated AST code generation tooling into active developer IDE environments, speeding up live type parsing and syntax coloring by 3.5x.
Maintained modular React and Next.js developer dashboards displaying complex syntax-tree hierarchies in real time.

TypeScriptRust/WASMNext.jsASTReact

Initialize Connection

Port 80/443 handshake & raw mailbox routing

shell_tunnel.sh

// Route packets directly to Kaustubh's mailbox

$ ping -c 1 goelkaustubh2006@gmail.com

64 bytes from goelkaustubh2006@gmail.com: icmp_seq=1 ttl=64 time=4.20 ms
--- connection handshake complete ---

Mailto: goelkaustubh2006@gmail.com

GitHub LinkedIn