Position: AI/ML Engineer (SDE - III)Location: HybridReporting to: CPO, Arré Voice
About Arré & Arré Voice
Arré is one of India’s leading digital media and media-tech companies, built at the intersection of storytelling, entertainment, and technology. Founded by media veterans, Arré grew rapidly into a multi-format studio known for its original web series, documentaries, podcasts, and social-first content that has reached more than 200 million viewers across India. Over the years, Arré has evolved from a premium content studio into a diversified media-tech ecosystem — a space where new formats, creators, and ideas are discovered, nurtured, and scaled across both traditional and emerging platforms.Arré Voice is the company’s flagship audio-first social platform — a short-form, voice-only space designed for India’s next billion users. Built around the idea that “your voice is your identity,” Arré Voice removes the pressure of video and makes self-expression radically simple. Users record 30-second “Voicepods,” participate in “Voicepools” (asynchronous audio threads), and engage inside “Voiceclubs” — vibrant interest-led communities where conversations flow naturally.With a creator network of 10,000+ voices and a monthly reach of 15–20 million across the Arré ecosystem, Arré Voice has become the destination for authentic, real, everyday storytelling from Tier-2 and Tier-3 India. The platform spans genres including personal storytelling, love & relationships, comedy, self-improvement, true crime reactions, Tamil and Hindi city-based discussions, and micro-podcasting formats that encourage participation over perfection.Arré Voice is also privacy-safe, multilingual, and built for low-pressure social interactions — enabling anyone to express themselves without being “camera ready.” The platform is expanding rapidly with creator tools powered by AI, emotion-aware voice workflows, and intuitive discovery systems that make it easy for new users to engage, create, and build their own voice-based identities.Together, Arré and Arré Voice represent India’s most ambitious push toward audio-native social storytelling — empowering millions to speak, listen, connect, and be heard.
About The Role
We are looking for an AI/ML Production Engineer with 4–5 years of experience building and deploying production-grade AI workflows at scale. This is a specialist engineering role for someone who has deep experience integrating foundation models (LLMs, TTS, STT, vision, speech, emotion detection, embeddings) and has worked extensively on real-time AI pipelines, scalable microservices, and backend infrastructure.You will work closely with Product, Backend, and Applied AI teams to enable next-generation experiences across speech, audio creation, Indic languages, and creator tools.What You’ll Work On
- Build and maintain production-ready AI inference pipelines for:
- Text-to-Speech (TTS)
- Speech-to-Text (STT)
- Emotion recognition
- LLM-based text workflows
- Image & video generation models
- Multimodal model integration
- Build backend services that orchestrate AI workflows end-to-end, including:
- Model routing
- Fallbacks and failovers
- Latency optimization
- Content safety filters
- Queues, batching, caching layers
- Integrate with Indic foundational models, including: Sarvam AI, Bhashini, Krutrim, Fireworks, Stability, OpenAI, Gemini, etc.
- Collaborate with product teams to convert voice-first use cases into scalable production systems.
- Own deployment pipelines end-to-end using:
- Docker
- Kubernetes
- Cloud GPU/CPU inference infra
- Monitoring & observability
- Work with audio teams on real-time features like:
- Creator Studio AI tools
- Journaling summaries
- Clipping & highlights generation
- Audio emotion tagging
- Auto-video generation
- Benchmark models, evaluate outputs, and recommend the best choices for cost, latency, and quality.
Must-Have Skills
AI/ML Production Expertise
- Hands-on experience deploying AI models in production (not research only).
- Strong understanding of TTS, STT, embeddings, RAG flows, LLM orchestration.
- Experience with multilingual/Indic language pipelines.
- Knowledge of model latency, throughput, caching, batching, token optimization.
Backend Engineering
- Strong in Python (preferred) or Node.js/Go.
- Deep knowledge of microservices, API design, async systems.
- Experience with Redis, queues (Kafka/RabbitMQ), PostgreSQL/Mongo.
Infrastructure
- Docker, Kubernetes, CI/CD pipelines.
- Hands-on experience with GPU workloads (NVIDIA stack, Triton, etc.).
- Familiarity with vector databases (Pinecone, Chroma, Weaviate, Milvus).
General Engineering
- Strong fundamentals in computer science (algorithms, data structures, concurrency).
- Experience debugging distributed systems in production.
- Ownership-driven mindset; ability to ship fast.
Bonus Skills
- Experience integrating model marketplaces (HuggingFace, Fireworks, Replicate).
- Familiarity with algorithmic audio or DSP pipelines.
- Applied experience with:
- Whisper / Speech models for Indic languages
- VLM models (Video/vision large models)
- Token-level emotion detection
- Experience with fine tuning small LLMs or TTS models for local languages.
- Prior experience in media, audio, creator platforms.
Who You Are
- A hands-on engineer who enjoys solving complex AI workflow problems.
- Obsessed with quality, latency, and production readiness.
- Comfortable operating in fast-moving, ambiguous environments.
- Excited about building the next generation of voice-first experiences at scale.
What We Offer
- Opportunity to build AI-first products loved by millions of creators.
- High ownership and autonomy.
- Work with new models, experimental architectures, and state-of-the-art tools.
- A culture that values speed, clarity, and craftsmanship.
Skills: ml,generative ai,llm models,ai workflows,foundational model,dataset preparation,model training
