Adaptive Audio Pipeline Creation

I need an audio processing expert (NOT a general developer) to build a fully adaptive audio pipeline.

Goal:
No matter the input (YouTube, Facebook, speech, music, noisy recordings), the output should always have a consistent sound:

- Warm
- Clear vocals (strong mid presence)
- Clean and controlled dynamics
- Pleasant listening experience

Technical requirements:

- MP3 output (~35–40 kbps VBR, LAME)
- Mono
- 44.1 kHz
- FFmpeg / Linux environment

Target metrics:

- Integrated loudness: ~ -16 LUFS (±1 LU)
- Loudness range (LRA): ~ 3–5 LU
- True peak: below -1.5 dBTP

Processing expectations:

- Adaptive pipeline (NOT static presets)
- Audio analysis (loudness, dynamics, frequency balance)
- Decision-based processing depending on input
- EQ + compression + filtering chain

Tonal shaping guidelines:

- Control low-end (avoid muddy bass <120 Hz)
- Add warmth (200–400 Hz)
- Enhance vocal clarity (2–4 kHz)
- Smooth highs (avoid harshness >10 kHz)

Important:
If you do not have real experience in audio processing, DSP, or sound engineering, please DO NOT apply.

Bonus if you worked with:

- FFmpeg audio filters
- Loudness normalization (LUFS)
- Psychoacoustics / low bitrate optimization

I can provide reference audio samples (original + processed).

Budget: $50–$150 (open to discussion based on experience and approach)

Looking for someone experienced in sound, not just coding.