NLP Engineer Multilingual AI (Bengaluru)

NLP Engineer Multilingual AI (Bengaluru)

19 Nov
|
BoloIndiaMediaNetworks
|
Bengaluru

19 Nov

BoloIndiaMediaNetworks

Bengaluru

Job Title: Lead – Multilingual AI & Language Technology / Indian Languages Engineering

Location:
India

Experience:
5–12+ years Employment: Full-time Domain: AI / NLP / Generative AI / Speech & Language Technologies / Social Media (PEOPLE WITHOUT NLP EXPERIENCE NEED NOT APPLY)

About the Role
We are looking for an experienced Multilingual AI & Language Technology expert to architect and build the language technology stack for Bolo India, enabling users across India to create, consume and interact with content in 50+ regional languages and dialects. The role will focus on automatic language detection, transliteration, translation, multilingual search, text normalization, language understanding and AI-powered content transformation. A key responsibility will be enabling a user to post content in their preferred Indian language and allowing other users to read and understand the same content in their preferred language, while preserving the original meaning, context, tone and cultural nuances.

Key Responsibilities
1. Multilingual Language Architecture
• Design the overall Indian Multilingual AI Architecture supporting 50+ languages/dialects.
• Develop language processing pipelines for: o Language identification o Script identification o Transliteration o Translation o Text normalization o Spell correction o Named Entity Recognition o Sentiment and emotion analysis o Multilingual semantic understanding
• Design a scalable language layer that can support addition of new languages without major application changes. 2. Auto-Transliteration Build intelligent transliteration capabilities such as: User types: "Aaj mausam bahut accha hai" System can automatically recognize the language/script and provide: आज मौसम बहुत अछा है Similarly, users should be able to type/read content across scripts such as:
• Hindi ↔ Devanagari
• Kannada ↔ Kannada script
• Tamil ↔ Tamil script
• Telugu ↔ Telugu script
• Malayalam ↔ Malayalam script
• Bengali ↔ Bengali script
• Gujarati ↔ Gujarati script
• Punjabi ↔ Gurmukhi
• Odia ↔ Odia
• Urdu ↔ Perso-Arabic while also supporting Romanized Indian language input. 3. Multilingual Translation Engine Develop/ integrate AI translation capabilities for:
• Post captions
• Comments
• Replies
• Hashtags
• User profiles
• Notifications
• UI text
• News/content
• Video subtitles
• Long-form content The system should intelligently understand the source language and translate content into the user's preferred language on demand. 4. Indian Language NLP Work on NLP models optimized specifically for Indian language characteristics, including:
• Code-mixing
• Code-switching
• Romanized Indian languages
• Regional slang
• Abbreviations
• Local expressions
• Spelling variations
• Dialects
• Informal social-media language
• Hinglish and other mixed-language content Example: "Namma Bengaluru weather super ide" The system should understand the mixed-language context rather than treating individual words independently. 5. AI/ML Model Development Develop, fine-tune and deploy models for:
• Multilingual Transformers
• LLMs
• Neural Machine Translation
• Indic language models
• Embedding models
• Sentence transformers
• Multilingual classification
• Retrieval-Augmented Generation
• Translation models
• Transliteration models




• Language identification models Experience with models/frameworks such as:
• IndicBERT
• IndicTrans / IndicTrans2
• mBERT
• XLM-R
• T5 / mT5 •Llama-family multilingual models
• Mistral-family models
• NLLB
• Whisper or equivalent speech models
• Custom Transformer architectures would be highly valuable. 6. Social Media Language Intelligence Because Bolo India is a social platform, the engineer must solve language problems in real-world user-generated content, not only clean textbook language. Build capabilities for:
• Multilingual content moderation
• Offensive-language detection
• Hate-speech detection
• Spam detection
• Toxicity detection
• Sentiment analysis
• Contextual understanding
• Multilingual keyword detection
• Safe-search classification
• Regional-language content recommendation The system should be able to understand the same abusive or harmful meaning expressed in different languages, scripts or Romanized text. 7. Auto Language Detection Develop a highly accurate language detection layer capable of identifying:
• Language
• Script
• Mixed-language content
• Romanized language
• Multiple languages within the same post For example: "Bangalore alli ivattu weather tumba chennagide" should be recognized as Kannada written in Roman script, rather than simply classified as English. 8. Transliteration API & Microservices Design production-grade APIs for: Input → Detection → Normalization → Transliteration → Translation → Output Requirements include:
• REST/gRPC APIs
• Low latency
• Horizontal scalability
• Caching
• Model serving
• Rate limiting
• API monitoring
• Failover
• Version management
• Model A/B testing 9. AI Infrastructure & Model Deployment Work closely with Cloud/DevOps teams to deploy multilingual AI models at scale. Experience with:
• AWS / Azure / GCP
• Kubernetes
• Docker
• GPU infrastructure
• GPU inference optimization
• NVIDIA CUDA
• TensorRT
• ONNX
• Hugging Face
• MLflow
• Model registries
• Distributed inference
• Batch and real-time inference is preferred. 10. Performance & Scalability The language platform must potentially process millions of posts, comments and interactions. The candidate will be responsible for designing systems capable of:
• High-volume inference
• Low-latency translation
• Real-time transliteration
• Asynchronous translation
• Queue-based processing
• Model caching
• GPU optimization
• Auto-scaling
• Fault tolerance Special attention should be given to cost per translation/inference as the platform scales. 11. Language Quality & Human-in-the-Loop Establish mechanisms for continuously improving language quality through:
• Human linguistic validation
• Native-language reviewers
• User feedback
• Translation confidence scores
• Error detection
• Model evaluation
• Regional language datasets
• Fine-tuning datasets




• Continuous learning pipelines The system should identify when AI is uncertain and avoid generating misleading translations. Required Technical Skills, Must Have:
• Strong NLP / Computational Linguistics background
• 5+ years in NLP/AI/ML engineering
• Strong Python
• PyTorch and/or TensorFlow
• Transformers
• Hugging Face ecosystem
• Multilingual NLP
• Neural Machine Translation
• Transliteration
• Language detection
• LLMs
• Embeddings and semantic search
• Model fine-tuning
• REST/gRPC APIs
• Production ML deployment
• Cloud and containerized environments Indian Language Expertise: Solid understanding of Indic language technology is highly desirable. Experience with multiple languages such as: Hindi, Bengali, Telugu, Marathi, Tamil, Gujarati, Urdu, Kannada, Malayalam, Odia, Punjabi, Assamese, Maithili, Sanskrit, Konkani, Nepali, Manipuri, Kashmiri, Sindhi and other Indian languages will be an advantage. Candidates who have worked with Indic NLP, AI4Bharat, Bhashini, CDAC or similar Indian-language technology ecosystems would be particularly valuable. AI Skills Required:
• Generative AI
• LLMs
• Prompt engineering
• LLM fine-tuning
• LoRA / PEFT
• RAG
• Multilingual embeddings
• Vector databases
• Transformer architectures
• NLP pipelines
• NMT
• Speech-to-text
• Text-to-speech
• OCR
• Multimodal AI
• AI model evaluation
• Model optimization
• Quantization
• GPU inference
• MLOps Advantage/Preferred: Experience in any of the following will be a major advantage:
• Social media platforms
• Google/Meta/Microsoft language technologies
• Indic language AI
• Search engines
• Content recommendation systems
• Translation platforms
• Voice assistants
• Speech AI
• Large-scale content moderation
• Generative AI platforms
• News/media platforms Key Deliverables The candidate will be expected to help build: 1. Indian Language Detection Engine Detect 50+ languages and mixed-language content. 2. Auto-Transliteration Engine Convert between Romanized and native scripts. 3. AI Translation Engine Translate user-generated content between Indian languages. 4. Multilingual NLP Engine Understand sentiment, intent, entities, toxicity and context. 5. Multilingual Moderation Engine Detect harmful content regardless of language/script. 6. Multilingual Search Engine Allow users to search content across languages. 7. Multilingual Recommendation Layer Recommend relevant content irrespective of the language in which it was created. 8. Language AI API Platform Provide reusable language APIs to all Bolo India applications and services. KPIs: The role will be measured on:
• Language detection accuracy
• Transliteration accuracy
• Translation quality/BLEU, COMET or equivalent metrics
• Latency per request
• AI inference cost
• Number of languages supported
• Percentage of content successfully processed
• Multilingual moderation accuracy
• Reduction in untranslated content
• Model uptime and reliability
• User language engagement
• Continuous improvement in language quality Ideal Candidate Profile We are looking for someone who is not merely an NLP developer, but someone capable of building an India-scale Language Intelligence Platform. The ideal candidate should be able to answer: "How do we enable users on Bolo India to communicate with each other digitally, regardless of the language or script they use?"

📌 NLP Engineer Multilingual AI (Bengaluru)
🏢 BoloIndiaMediaNetworks
📍 Bengaluru

Reply to this offer

Impress this employer describing Your skills and abilities, fill out the form below and leave Your personal touch in the presentation letter.

Subscribe to this job alert:

Get the latest job offers by email for: nlp engineer multilingual ai (bengaluru) / bengaluru

Subscribe to this job alert:

Get the latest job offers by email for: nlp engineer multilingual ai (bengaluru) / bengaluru