Founding Engineer (Voice AI — Call Transcription & LLM Systems) – Internship (WFH)
Investcode Private Limited is recruiting a Founding Engineer intern for Voice AI, Call Transcription and LLM Systems. Although the role accepts freshers or eighth-semester engineering candidates, it asks for hands-on technical experience with speech systems. The internship is remote, lasts three months and advertises three openings at ₹15,000–25,000 per month.
Internship at a glance
| Employer | Investcode Private Limited |
|---|---|
| Work mode | Work from home |
| Duration | 3 months |
| Stipend | ₹15,000–25,000 per month |
| Openings | 3 |
| Start window | 2 November–7 December 2026 |
| Apply by | 2 December 2026 |
What you will work on
- Build and improve voice-to-text pipelines for real calls, including challenging accents, interruptions, low-quality audio and domain-specific language.
- Use language models to understand, summarise and extract insights from transcripts.
- Design speaker diarisation, call summarisation and intent-extraction systems, with real-time or near-real-time transcription.
- Improve production latency, accuracy and cost while moving models from experiments into working systems.
Eligibility and availability
Freshers or eighth-semester BE/BTech candidates in Computer Science or a related technical field, with hands-on voice/speech LLM and ASR experience, strong Python, ML/NLP and system-design fundamentals, and three-month WFH availability.
Candidates must be available to start between 2 November–7 December 2026 and commit to the advertised duration. The application deadline and the joining window serve different purposes: apply by 2 December 2026 even when the allowed joining window ends later.
Skills and fit
The required skill labels are machine learning, natural language processing and LLMOps. Other requirements include experience with ASR tools such as Whisper, Deepgram, Vosk or proprietary systems, transcript post-processing with LLMs, speaker diarisation, timestamp alignment, chunking and streaming. Strong Python and experience with real call-transcription use cases are explicitly requested. Real-time transcription, domain-specific audio, inference optimisation and relevant open-source contributions are bonuses.
Stipend and listed benefits
The stipend is ₹15,000–25,000 per month, with no separate incentive split disclosed. The listing offers a recommendation letter, flexible hours and five working days a week. A possible permanent conversion carries an advertised ₹6–15 lakh annual range; it is conditional on successful conversion and is not the internship stipend.
About the work
The employer describes InvestKode AI as a small team building equities-research products with language models, real-time data and voice intelligence. Applicants should be comfortable with the stated production focus, rather than assuming that the role is limited to classroom machine-learning exercises.
How to apply
- Open the exact Internshala listing through the application link below and read the current role details.
- Sign in or create your candidate account when Internshala asks you to continue. Keep your profile, relevant skills and availability accurate.
- Use Apply now for this specific opening and complete any employer questions shown in the application. Explain the experience relevant to the duties above.
- Submit before 2 December 2026 and check your candidate account for further communication. An application is not a guarantee of selection.
Before applying, confirm that the duration, work mode and pay terms fit your circumstances. Where working-hour details or payment conditions are not specified, ask the employer during the hiring process rather than assuming them.
