Internships for B.Tech and B.Sc Horticulture
YuktYantra provides structured, mentored undergraduate research internships and capstone fellowships for students pursuing B.Tech and B.Sc Horticulture degrees. Work on physical perception, agricultural camera carts, and conversational speech models.
Real industrial data, clear evaluation, professional mentorship
Undergraduate students rarely get to work on real, uncurated physical and agronomic data before graduation. Our internship programme bridges that gap.
Travel Costs Covered (Unpaid Research)
Local bus transit expenses reimbursed for scheduled vineyard field sessions. This is an unpaid research internship and academic capstone project fellowship.
Direct Mentorship by Arun Patil
Regular architecture and code reviews led directly by Arun Patil (17+ years in machine learning systems, specialized AI compilers, and retail computer vision platforms).
Certificate and Reference Letter
Formal internship certificate and detailed recommendation letter from Arun Patil documenting specific technical achievements. Contributor attribution on open public release notes.
Programme Structure & Selection Steps
Key administrative parameters for prospective student researchers and faculty departmental advisors.
Flexible around university semester exam calendars.
Part-time during semester; 25–30h during term breaks.
Small, close-knit cohort with high-touch 1-on-1 reviews.
Rolling intake aligned with winter and spring university terms.
Four-Step Selection Process:
- Application Review: Email resume, academic transcript, and relevant project links (GitHub or coursework portfolio).
- Technical Discussion: 30-minute technical problem walkthrough with Arun Patil on perception or systems logic.
- Agreement Execution: Standard 1-page student IP assignment and NDA signed prior to project onboarding.
- Cohort Kickoff: Repository access, initial dataset samples, and bi-weekly milestone roadmap issued.
Three Structured Internship Tracks
Each project track maps onto an ongoing YuktYantra applied research initiative with defined student deliverables.
Mobile Challenge Pose & Success Rater
Implement on-device computer vision models to score camera-based household manipulation challenges: Steady Stack (tumbler verticality and wobble), Dal Sort (bean bounding boxes and bowl count purity), and Fold Off (cloth edge parallelism).
- • Hand pose estimation (MediaPipe / custom PyTorch)
- • End-state segmentation & scoring heuristic design
- • Benchmark validation dataset with failure modes logged
- • Recommended for: B.Tech CSE, ISE, AI/ML students
Viticulture Smartphone Imaging Cart & Field Survey
Assemble and operate a lightweight manual push-cart rig using mounted smartphones across vineyard rows in Sindagi taluk. Record video passes and collect physical hand measurements (berry sizing with vernier calipers and refractometer sugar readings) across phenological stages to build an initial open dataset.
- • Smartphone video acquisition & spatial row indexing
- • Hand physiological measurements (caliper diameter & Brix refractometer)
- • Scheduled field visits in Sindagi taluk (Vijayapura district)
- • Recommended for: B.Sc Horticulture, B.Tech Agriculture, ECE, Mechanical, CSE
Full-Duplex Speech & Indic Telephony Adapters
Profile open full-duplex speech foundation models (PersonaPlex, Human-1) for Indian telephony. Full-scale GPU profiling and model fine-tuning will run when GPU grant allocations or compute sponsorship are activated. Current prep work focuses on CPU inference tracing, 8 kHz narrowband audio normalizers, and synthetic Kannada dialogue dataset curation.
- • PyTorch CPU inference profiling & execution tracing
- • 8kHz vs 24kHz audio tokenization quality benchmarks
- • Synthetic two-channel conversational Kannada dataset curation
- • Note: GPU benchmarking scheduled upon compute cluster allocation
Project Agreement & Attribution
All project source code, models, trade secrets, and field datasets remain the exclusive intellectual property of Arun Patil Labs. Interns sign a clean 1-page student NDA and IP assignment agreement upon acceptance.
Student Credit: Active student contributors receive formal contributor attribution in public dataset cards and release documentation, along with formal completion certificates and detailed recommendation letters.
Vineyard Field Visit Guidelines
- • Location: Field visits take place at our private research vineyard block in Sindagi taluk, Vijayapura district (exact directions provided upon selection).
- • Transport: Local public bus fare is reimbursed against tickets for authorized study sessions.
- • Consent: Accepted students must provide signed parental consent and college no-objection/acknowledgement forms.
- • Insurance: Students must have active student health/accident insurance.
- • Safety Gear: Farm visits require standard outdoor protection (sun hat, closed footwear, and hydration).
How to Apply
Open to 2nd, 3rd, and 4th year students pursuing B.Tech and B.Sc Horticulture degrees. We welcome independent student applicants as well as college-sponsored capstone teams.
Faculty advisors interested in joint capstone supervision or student recommendations are invited to connect directly with Arun Patil.
Initiate Application Inquiry
Send student resume, GitHub/portfolio link, and preferred project track to:
YuktYantra · A unit of Arun Patil Labs · Bengaluru & Sindagi Taluk