AI Engineer

Company Description

We are a US VC-backed AI film studio with major Bollywood and Tollywood projects underway and a clear path into Hollywood. With multiple full-length films, music videos, and commercial projects in the pipeline, we’re building the world’s first end-to-end AI-powered film studio. 

ROLE 

• Full-time, on-site in Hyderabad. We’re hunting a cracked AI engineer to wire state-of-the-art models straight into real productions. Your users aren’t PMs—they’re A-list directors, actors, DoPs, and post houses. One day it’s a diffusion pipeline for shot-matching; the next it’s a GPU render queue that saves a cut before midnight. Expect edit-bay drop-ins, set visits, and “need this tonight” pings from big names. Own problems idea → shipped → iterated in hours, not quarters. No degree required; high-school candidates welcome if you can show work.

What you’ll do

• Build + ship video/image/audio AI pipelines (generation, cleanup, upscaling, relighting, transcription, lip-sync)
• Solve temporal consistency, shot/character continuity, matting/rotoscoping, depth/normal fusion
• Optimize inference: PyTorch, mixed precision (FP16/INT8), batching, TensorRT/ONNX, multi-GPU scheduling
• Serve models in prod: Triton/Ray/Celery workers, queues, webhooks, artifact lineage
• Fine-tune with LoRA/QLoRA, dataset curation, evals; track with W&B
• Glue to post: FFmpeg, frame-accurate I/O, ProRes/EXR, basic ACES color sanity
• Pair with full-stack to build review/approval + render queues that artists actually use
• Used the best video AI models out there — ideally shipped with some of: Runway Gen-3, Pika, Luma Dream Machine, Kling, Google Veo, Stable Video Diffusion (SVD) / AnimateDiff/ControlNet; for cleanup: SAM/HQ-SAM, RVM, XMem, GFPGAN/CodeFormer; for lip-sync: Wav2Lip/SadTalker.

Compensation

₹2,25,000 – ₹2,75,000 per month

Qualification: We need a cracked AI guy even if you’re in high school. No degree/pedigree needed; show us what you’ve shipped.