Search by job, company or skills

Senior / Principal ML Engineer, Social Listening

Senior / Principal ML Engineer, Social Listening

Indian Institute Of Science
5-7 Years
Not Disclosed
  • Posted 31 minutes ago
  • Be among the first 10 applicants

Job Description

Senior / Principal ML Engineer, Social Listening

SPIRE Lab (https://spire.ee.iisc.ac.in/), Indian Institute of Science (IISc), Bengaluru · Full-time at IISc · 5+ years of experience

About Social Listening

SPIRE Lab at IISc is hiring a senior individual contributor to build the AI, data and evaluation backbone of a two-year, Gates Foundation-funded research programme in rural Bihar.

The programme asks a simple question: can voice AI help a state government genuinely listen to rural women Women in self-help groups (Didis) have views on livelihoods, nutrition and local services, but little of it reaches decision makers in a usable form.

In partnership with the Bihar Rural Livelihoods Promotion Society (BRLPS / JEEViKA), we will build and field-test a voice AI system that lets SHG women and field cadres ask questions, share feedback and raise grievances over basic and shared phones, in Hindi, Bhojpuri and Magahi. A sense-making layer turns those calls into structured insights JEEViKA can act on. Alongside it, we will build an open evaluation sandbox that tests how voice AI actually performs in the field, with rural accents and real phone lines, rather than only in the lab. The baseline solution, evaluation framework and technical guidelines are intended to be published so other programmes can reuse them.

A field implementation partner will run the voice platform and field operations under IISc's technical oversight. This role owns the AI, data and evaluation work on the IISc side.

Gender-intentional design runs through the work.

About the team

You will work inside SPIRE Lab as a senior individual contributor. You set technical direction through the quality of your work, not by managing a team.

You will sit with speech and AI researchers at IISc, work closely with the Head of Program and a Program Manager who owns operations and partner coordination, and collaborate with government, civil society and research partners on field operations, data governance and evaluation.

About IISc / SPIRE Lab

The Signal Processing, Interpretation and REpresentation Laboratory (SPIRE Lab) is part of the Department of Electrical Engineering at the Indian Institute of Science (IISc), Bengaluru. Founded in May 2015 and directed by Dr. Prasanta Kumar Ghosh, the lab works on signal processing, speech processing, representation learning and machine learning, with a focus on human-centred signals such as speech and audio.

The lab publishes widely, works with government, industry and civil society partners, and is releasing public datasets for Indian accents and languages. It is also a pioneer of speech datasets for Indian languages, having began work on datasets in 2020, followed by founding Project Vaani in collaboration with Google Deepmind in 2023. It actively works on speech datasets, model development and model evaluations, apart from smaller projects focussed on atypical speech.

This role is an IISc position within SPIRE Lab.

The role

You are the senior technical owner of the social listening machinery: the pipeline that turns raw call recordings into trustworthy, structured, actionable signal, and the evaluation that tells us whether it works for the women it is meant to serve.

You will be IISc's technical counterpart to the implementation partner, and you will ship work that can be published and reused. This is a hands-on individual contributor role.

What you'll do

  • Design and own the end-to-end pipeline: ingestion of call recordings and metadata from the partner's voice platform, speech recognition for Bihar's languages and dialects, translation, classification, theme and grievance extraction, and insight reporting.
  • Build the evaluation program for voice AI in real conditions: noisy shared feature phones, code-mixing, and low-resource dialects. Define metrics with research collaborators, build eval datasets and harnesses, and report honestly on where models fail and for whom, including gender and subgroup analysis.
  • Be IISc's technical counterpart to the implementation partner: review platform architecture, data export and integration interfaces, consent and data-protection practices, and service levels, so the platform meets research and evaluation needs.
  • Design the human-in-the-loop layer with the cadre contextualization and case-management workflow, so AI output is checked and acted on, not only dash-boarded.
  • Own data governance in practice: consent, de-identification, secure storage and access, aligned with India's DPDP Act.
  • Ship open technical frameworks, documentation and implementation guidelines so other programmes can replicate the work.
  • Write clearly for government, funder and research audiences. Work in the field when needed.

What we're looking for

  • At least 5 years building and shipping ML or speech systems in production, with real ownership of a pipeline end to end.
  • Strong computer science fundamentals, and fluency in Python. This role is hands-on.
  • Experience with, or a demonstrated interest in, voice-based AI for Indic languages. Judgement to choose between open models, fine-tuning and vendor APIs.
  • Strong evaluation instincts: you have built eval sets and harnesses, and you distrust a single headline metric.
  • Comfort with messy real-world data, and with working alongside field teams and government partners.
  • Clear writing, first-principles thinking, and the ability to lead technically without being managed closely. You are an individual contributor, not a people manager.

Nice to have

  • Hands-on work with ASR, TTS or real-time voice agents, especially in low-resource or code-mixed settings.
  • Experience with social listening, IVR or voice platforms, feedback or grievance systems, or public-health or livelihoods programmes in rural India.
  • Familiarity with gender-intentional design, survey methodology or social-science collaborators.
  • Hindi, plus Bhojpuri, Magahi or Maithili.
  • An open-source or publication track record.
  • Comfort with cloud infrastructure (AWS, GCP or Azure).

Contract & logistics

Location: SPIRE Lab, IISc Bengaluru, with travel to Bihar

Type: Full-time IISc

How to apply

[Apply via this Form. (https://forms.gle/3yArL6GrkBxKFfph8) Applications are reviewed on a rolling basis.]

More Info

Job Type:
Industry:
Employment Type:

Key Skills

Similar Jobs

5-7 yrs
Bengaluru, India
Skills:
lifecycle management , Large Language Models (LLMs), Machine Learning, Scalability, Python, Transformers, GenAI Technologies, AI Agents, Model Deployment, Prompt Engineering, MLOps Practices, Observability