prepare for huggingface deployment

.dockerignore

@@ -0,0 +1,34 @@
+# Environment
+.env
+
+# Python cache
+__pycache__/
+*.pyc
+*.pyo
+*.pyd
+
+# Virtual environments
+venv/
+.venv/
+env/
+inter/
+
+# Node / frontend (not needed for backend deploy)
+frontend/
+node_modules/
+
+# Uploads (runtime data)
+uploads/
+
+# Git
+.git/
+.gitignore
+
+# OS / IDE junk
+.DS_Store
+Thumbs.db
+.idea/
+.vscode/
+*.swp
+
+backend/.env

.gitignore

@@ -15,3 +15,4 @@ WORKFLOW.md
 voice_name_list_xtts.txt
 
 
+SYSTEM_ARCHITECTURE.md

Dockerfile

@@ -0,0 +1,28 @@
+FROM python:3.10-slim
+
+ENV PYTHONUNBUFFERED=1
+ENV PYTHONDONTWRITEBYTECODE=1
+
+WORKDIR /app
+
+RUN apt-get update && apt-get install -y \
+    gcc \
+    ffmpeg \
+    libglib2.0-0 \
+    libsm6 \
+    libxext6 \
+    libxrender-dev \
+    && rm -rf /var/lib/apt/lists/*
+
+# Copy ONLY backend requirements
+COPY backend/requirements.txt .
+
+RUN pip install --no-cache-dir --upgrade pip \
+    && pip install --no-cache-dir -r requirements.txt
+
+# Copy backend code into container
+COPY backend/ .
+
+EXPOSE 7860
+
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

backend/.dockerignore

@@ -1,73 +0,0 @@
-# Environments
-.env
-
-# Byte-compiled / optimized / DLL files
-__pycache__/
-*.py[cod]
-*$py.class
-
-# Virtual environments
-venv/
-.venv/
-env/
-.env/
-
-# Distribution / packaging
-build/
-develop-eggs/
-dist/
-downloads/
-eggs/
-.eggs/
-lib/
-lib64/
-parts/
-sdist/
-var/
-wheels/
-share/python-wheels/
-*.egg-info/
-.installed.cfg
-*.egg
-
-# PyInstaller
-*.manifest
-*.spec
-
-# Installer logs
-pip-log.txt
-pip-delete-this-directory.txt
-
-# Unit test / coverage reports
-htmlcov/
-.tox/
-.nox/
-.coverage
-.coverage.*
-.cache
-nosetests.xml
-coverage.xml
-*.cover
-*.py,cover
-.hypothesis/
-.pytest_cache/
-cover/
-
-# IDEs
-.idea/
-.vscode/
-*.swp
-*.swo
-
-# OS generated files
-.DS_Store
-.DS_Store?
-._*
-.Spotlight-V100
-.Trashes
-ehthumbs.db
-Thumbs.db
-
-# Git
-.git/
-.gitignore

backend/Dockerfile

@@ -1,24 +0,0 @@
-# Use the official Python base image
-FROM python:3.10-slim
-
-# Set environment variables
-ENV PYTHONDONTWRITEBYTECODE=1
-ENV PYTHONUNBUFFERED=1
-
-# Set the working directory in the container
-WORKDIR /app
-
-# Copy the requirements file into the container
-COPY requirements.txt .
-
-# Install dependencies
-RUN pip install --no-cache-dir -r requirements.txt
-
-# Copy the rest of the application code
-COPY . .
-
-# Expose port 8000
-EXPOSE 8000
-
-# Command to run the application
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000"]

backend/routers/admin.py

@@ -563,7 +563,66 @@ from services.chatbot_service import (
     process_chatbot_query,
     update_student_info,
     generate_excel,
+    filter_students_structured,
 )
+from schemas.admin import StudentFilterRequest, StudentFilterExportRequest
+
+
+@router.post("/students/filter")
+async def students_filter(
+    request: StudentFilterRequest,
+    current_user: dict = Depends(require_role("admin")),
+):
+    """Structured student filter — no AI, direct params (sort + date range + multi-test)."""
+    try:
+        result = await filter_students_structured(
+            group_test_ids=request.group_test_ids,
+            jd_id=request.jd_id,
+            start_date=request.start_date,
+            end_date=request.end_date,
+            top_k=request.top_k,
+            min_score=request.min_score,
+            sort_fields=request.sort_fields,
+            sort_orders=request.sort_orders,
+        )
+        return result
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+
+@router.post("/students/export-excel")
+async def students_export_excel(
+    request: StudentFilterExportRequest,
+    current_user: dict = Depends(require_role("admin")),
+):
+    """Generate styled Excel for structured student filter results."""
+    try:
+        bio = generate_excel(
+            rows=request.rows,
+            topic_columns=request.topic_columns,
+            group_test_name=request.group_test_name,
+        )
+        safe_name = request.group_test_name.replace(" ", "_").replace("/", "-")[:40]
+        filename = f"{safe_name}_students.xlsx"
+        return StreamingResponse(
+            bio,
+            media_type="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+            headers={"Content-Disposition": f'attachment; filename="{filename}"'},
+        )
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+
+@router.patch("/students")
+async def update_student(
+    request: ChatbotStudentUpdate,
+    current_user: dict = Depends(require_role("admin")),
+):
+    """Admin corrects a student's reg_no or name (shared by both filter endpoints)."""
+    try:
+        return await update_student_info(request.user_id, request.reg_no, request.name)
+    except ValueError as e:
+        raise HTTPException(status_code=400, detail=str(e))
 
 
 @router.post("/chatbot/query")

backend/schemas/admin.py

@@ -111,7 +111,7 @@ class RoleRequirementResponse(BaseModel):
     level: str
 
 
-# ── Chatbot ───────────────────────────────────────────────────────────────────
+# ── Chatbot (legacy) ──────────────────────────────────────────────────────────
 
 class ChatbotQueryRequest(BaseModel):
     query: str
@@ -128,3 +128,22 @@ class ChatbotStudentUpdate(BaseModel):
     user_id: str
     reg_no: Optional[str] = None
     name: Optional[str] = None
+
+
+# ── Structured Student Filter ─────────────────────────────────────────────────
+
+class StudentFilterRequest(BaseModel):
+    group_test_ids: Optional[List[str]] = None   # None = all tests
+    jd_id: Optional[str] = None
+    start_date: Optional[str] = None             # "YYYY-MM-DD"
+    end_date: Optional[str] = None               # "YYYY-MM-DD"
+    top_k: Optional[int] = None
+    min_score: Optional[float] = None
+    sort_fields: List[str] = ["time"]            # ordered priority: "time"|"score"|"duration"
+    sort_orders: List[str] = ["desc"]            # matching orders: "asc"|"desc"
+
+
+class StudentFilterExportRequest(BaseModel):
+    rows: List[dict]
+    topic_columns: List[dict]
+    group_test_name: str

backend/services/chatbot_service.py

@@ -3,6 +3,7 @@
 import json
 import re
 from collections import defaultdict
+from datetime import datetime, timezone, timedelta
 from io import BytesIO
 
 from bson import ObjectId
@@ -403,3 +404,263 @@ def generate_excel(rows: list[dict], topic_columns: list[dict], group_test_name:
     wb.save(bio)
     bio.seek(0)
     return bio
+
+
+# ─── Structured Student Filter (no Gemini) ───────────────────────────────────
+
+def _parse_date(date_str: str | None, end_of_day: bool = False):
+    """Parse YYYY-MM-DD string to UTC-aware datetime, or None."""
+    if not date_str:
+        return None
+    try:
+        dt = datetime.strptime(date_str.strip(), "%Y-%m-%d").replace(tzinfo=timezone.utc)
+        if end_of_day:
+            dt = dt + timedelta(days=1)
+        return dt
+    except ValueError:
+        return None
+
+
+def _compute_duration(started_at, completed_at) -> float | None:
+    """Duration in minutes, rounded to 1 dp. Returns None if either timestamp missing."""
+    if started_at is None or completed_at is None:
+        return None
+    try:
+        # Handle both datetime objects and ISO strings
+        if isinstance(started_at, str):
+            started_at = datetime.fromisoformat(started_at.replace("Z", "+00:00"))
+        if isinstance(completed_at, str):
+            completed_at = datetime.fromisoformat(completed_at.replace("Z", "+00:00"))
+        delta = completed_at - started_at
+        return round(max(delta.total_seconds(), 0) / 60, 1)
+    except Exception:
+        return None
+
+
+async def filter_students_structured(
+    group_test_ids: list[str] | None,
+    jd_id: str | None,
+    start_date: str | None,
+    end_date: str | None,
+    top_k: int | None,
+    min_score: float | None,
+    sort_fields: list[str] | None = None,
+    sort_orders: list[str] | None = None,
+) -> dict:
+    """Structured student filter — no Gemini, explicit params, composable."""
+    db = get_db()
+
+    # ── Fetch all group tests for metadata ────────────────────────────────────
+    gt_cursor = db[GROUP_TESTS].find({}).sort("created_at", -1)
+    gt_docs = await gt_cursor.to_list(length=300)
+    all_group_tests: dict[str, dict] = {
+        str(d["_id"]): {
+            "id": str(d["_id"]),
+            "name": d.get("name", ""),
+            "topic_ids": d.get("topic_ids") or [],
+        }
+        for d in gt_docs
+    }
+
+    # ── Fetch JD info if provided ─────────────────────────────────────────────
+    jd_content, jd_req_skills = (None, [])
+    if jd_id:
+        jd_content, jd_req_skills = await _jd_skills(jd_id, db)
+    use_jd_ranking = bool(jd_req_skills)
+
+    # ── Build MongoDB query ───────────────────────────────────────────────────
+    results_filter: dict = {}
+
+    if group_test_ids:
+        results_filter["group_test_id"] = {"$in": group_test_ids}
+
+    start_dt = _parse_date(start_date, end_of_day=False)
+    end_dt = _parse_date(end_date, end_of_day=True)
+    if start_dt or end_dt:
+        date_filter: dict = {}
+        if start_dt:
+            date_filter["$gte"] = start_dt
+        if end_dt:
+            date_filter["$lt"] = end_dt
+        results_filter["started_at"] = date_filter
+
+    results_cursor = db[GROUP_TEST_RESULTS].find(results_filter)
+    results_docs = await results_cursor.to_list(length=3000)
+
+    # ── Collect topic columns from selected group tests ───────────────────────
+    selected_gt_ids: set[str] = (
+        set(group_test_ids) if group_test_ids
+        else {str(d["_id"]) for d in gt_docs}
+    )
+
+    topic_columns: list[dict] = []
+    topic_seen: set[str] = set()
+    group_test_name = "All Group Tests"
+
+    if group_test_ids and len(group_test_ids) == 1:
+        gt = all_group_tests.get(group_test_ids[0])
+        if gt:
+            group_test_name = gt["name"]
+            for tid in gt["topic_ids"]:
+                if tid not in topic_seen:
+                    topic_seen.add(tid)
+                    try:
+                        t = await db[TOPICS].find_one({"_id": ObjectId(tid)})
+                    except Exception:
+                        t = None
+                    if t:
+                        topic_columns.append({"id": tid, "name": t.get("name", tid)})
+    elif group_test_ids and len(group_test_ids) > 1:
+        names = [all_group_tests[gid]["name"] for gid in group_test_ids if gid in all_group_tests]
+        group_test_name = ", ".join(names) if names else "Multiple Tests"
+        for gid in group_test_ids:
+            gt = all_group_tests.get(gid)
+            if not gt:
+                continue
+            for tid in gt["topic_ids"]:
+                if tid not in topic_seen:
+                    topic_seen.add(tid)
+                    try:
+                        t = await db[TOPICS].find_one({"_id": ObjectId(tid)})
+                    except Exception:
+                        t = None
+                    if t:
+                        topic_columns.append({"id": tid, "name": t.get("name", tid)})
+
+    # ── Group by (user_id, group_test_id) → pick best attempt ────────────────
+    attempts_map: dict[tuple, list] = defaultdict(list)
+    for r in results_docs:
+        uid = r.get("user_id", "")
+        gtid = r.get("group_test_id", "")
+        if uid:
+            attempts_map[(uid, gtid)].append(r)
+
+    rows: list[dict] = []
+
+    for (uid, gt_id), attempts in attempts_map.items():
+        best = None
+        for attempt in attempts:
+            attempt = await _refresh_topic_statuses(attempt, db)
+            score = attempt.get("overall_score") or 0
+            if best is None or score > (best.get("overall_score") or 0):
+                best = attempt
+
+        user = await _user_info(uid, db)
+
+        # Per-topic scores
+        topic_scores: dict[str, dict] = {}
+        for tr in best.get("topic_results") or []:
+            tid = tr.get("topic_id", "")
+            topic_scores[tid] = {
+                "topic_name": tr.get("topic_name", ""),
+                "score": tr.get("overall_score"),
+                "status": tr.get("status", "pending"),
+            }
+
+        # JD skill match
+        skill_match: float | None = None
+        if use_jd_ranking:
+            skills_doc = await db[SKILLS].find_one({"user_id": uid})
+            student_skills = (skills_doc or {}).get("skills") or []
+            skill_match = _skill_match_pct(student_skills, jd_req_skills)
+
+        # Attempt time & duration
+        started_at = best.get("started_at")
+        completed_at = best.get("completed_at")
+        attempt_time: str | None = None
+        if started_at is not None:
+            try:
+                attempt_time = started_at.isoformat() if isinstance(started_at, datetime) else str(started_at)
+            except Exception:
+                attempt_time = None
+
+        # Collect topic columns dynamically when showing all tests
+        if not group_test_ids:
+            gt = all_group_tests.get(gt_id)
+            if gt:
+                for tid in gt["topic_ids"]:
+                    if tid not in topic_seen:
+                        topic_seen.add(tid)
+                        try:
+                            t = await db[TOPICS].find_one({"_id": ObjectId(tid)})
+                        except Exception:
+                            t = None
+                        if t:
+                            topic_columns.append({"id": tid, "name": t.get("name", tid)})
+
+        gt_info = all_group_tests.get(gt_id, {})
+
+        row = {
+            "user_id": uid,
+            "reg_no": user["reg_no"],
+            "name": user["name"],
+            "email": user["email"],
+            "group_test_id": gt_id,
+            "group_test_name": best.get("group_test_name") or gt_info.get("name", ""),
+            "overall_score": round(best.get("overall_score") or 0, 1),
+            "total_attempts": len(attempts),
+            "status": best.get("status", "in_progress"),
+            "topic_scores": topic_scores,
+            "skill_match": skill_match,
+            "rank": 0,
+            "attempt_time": attempt_time,
+            "duration_minutes": _compute_duration(started_at, completed_at),
+        }
+        rows.append(row)
+
+    # ── Min score filter ──────────────────────────────────────────────────────
+    if min_score is not None:
+        rows = [r for r in rows if (r["overall_score"] or 0) >= min_score]
+
+    # ── Multi-sort ────────────────────────────────────────────────────────────
+    _fields = sort_fields if sort_fields else ["time"]
+    _orders = sort_orders if sort_orders else ["desc"]
+    _INF = float("inf")
+    _NEG_INF = float("-inf")
+
+    def _row_key(r: dict, field: str, order: str):
+        """Return a comparable key, handling None with order-aware sentinel."""
+        desc = order.lower() == "desc"
+        if field == "score":
+            v = r.get("overall_score") or 0
+            return -v if desc else v
+        elif field == "duration":
+            v = r.get("duration_minutes")
+            if v is None:
+                return _INF  # always sort None to end
+            return -v if desc else v
+        else:  # "time"
+            v = r.get("attempt_time") or ""
+            # For strings, desc means we negate via reverse tuple trick below
+            return v
+
+    # Apply sorts in reverse priority order (stable sort)
+    paired = list(zip(_fields, _orders))
+    for field, order in reversed(paired):
+        desc = order.lower() == "desc"
+        if field == "time":
+            rows.sort(key=lambda r: r.get("attempt_time") or "", reverse=desc)
+        elif field == "score":
+            rows.sort(key=lambda r: r.get("overall_score") or 0, reverse=desc)
+        elif field == "duration":
+            rows.sort(
+                key=lambda r: r["duration_minutes"] if r["duration_minutes"] is not None
+                else (_NEG_INF if desc else _INF),
+                reverse=desc,
+            )
+
+    # ── Assign ranks ──────────────────────────────────────────────────────────
+    for i, row in enumerate(rows):
+        row["rank"] = i + 1
+
+    # ── Top-K slice ───────────────────────────────────────────────────────────
+    if top_k and top_k > 0:
+        rows = rows[:top_k]
+
+    return {
+        "group_test_name": group_test_name,
+        "group_test_id": group_test_ids[0] if group_test_ids and len(group_test_ids) == 1 else None,
+        "topic_columns": topic_columns,
+        "rows": rows,
+        "total": len(rows),
+    }

resume-jd-verification-2026-04-10T05-15-44-248Z.pdf

@@ -1,646 +0,0 @@
-%PDF-1.3
-%�߬�
-3 0 obj
-<</Type /Page
-/Parent 1 0 R
-/Resources 2 0 R
-/MediaBox [0 0 595.2799999999999727 841.8899999999999864]
-/Contents 4 0 R
->>
-endobj
-4 0 obj
-<<
-/Length 5607
->>
-stream
-0.200025 w
-0 G
-BT
-/F2 16 Tf
-18.3999999999999986 TL
-0 g
-40. 795.8899999999999864 Td
-(Resume vs Job Description Verification) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 771.8899999999999864 Td
-(Verification ID:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 771.8899999999999864 Td
-(b5519c4e-2ab1-4bbd-a5ca-ff133b558b5b) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 753.8899999999999864 Td
-(Saved At:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 753.8899999999999864 Td
-(4/10/2026, 10:45:29 AM) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 735.8899999999999864 Td
-(Role:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 735.8899999999999864 Td
-(Generative AI Engineer) Tj
-ET
-BT
-/F2 16 Tf
-18.3999999999999986 TL
-0 g
-40. 711.8899999999999864 Td
-(Job Description Snapshot) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 687.8899999999999864 Td
-(JD Title:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 687.8899999999999864 Td
-(AI Engineering Intern) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 669.8899999999999864 Td
-(Company:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 669.8899999999999864 Td
-(-) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 651.8899999999999864 Td
-(Required Skills:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 651.8899999999999864 Td
-(Basic understanding of Machine Learning concepts \(supervised/unsupervised learning\),) Tj
-T* (Familiarity with Python and libraries like NumPy, Pandas, Scikit-learn, Knowledge of) Tj
-T* (deep learning frameworks \(e.g., TensorFlow or PyTorch\) is a plus Strong analytical and) Tj
-T* (problem-solving skills) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 595.8899999999999864 Td
-(JD Description:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 595.8899999999999864 Td
-(Key Responsibilities) Tj
-T* () Tj
-T* (Assist in developing and implementing AI/ML models and algorithms) Tj
-T* (Work on data preprocessing, cleaning, and analysis) Tj
-T* (Support model training, evaluation, and optimization) Tj
-T* (Conduct research on the latest AI trends and technologies) Tj
-T* (Collaborate with engineers and product teams to integrate AI solutions) Tj
-T* (Document experiments, processes, and results) Tj
-T* (Participate in brainstorming and problem-solving sessions) Tj
-ET
-BT
-/F2 16 Tf
-18.3999999999999986 TL
-0 g
-40. 463.8899999999999864 Td
-(Resume Snapshot) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 439.8899999999999864 Td
-(Resume File:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 439.8899999999999864 Td
-(Resume.pdf) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 421.8899999999999864 Td
-(Candidate:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 421.8899999999999864 Td
-(SAJITH J) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 403.8899999999999864 Td
-(Email:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 403.8899999999999864 Td
-(jsajith76@gmail.com) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 385.8899999999999864 Td
-(Phone:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 385.8899999999999864 Td
-(+91 8637440071) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 367.8899999999999864 Td
-(Location:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 367.8899999999999864 Td
-(Coimbatore, India) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 349.8899999999999864 Td
-(Extracted Skills:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 349.8899999999999864 Td
-(Python, SQL, RAG Pipelines, Semantic Search, Embedding Models, Vector Similarity) Tj
-T* (Search, Prompt Engineering, LangChain, LangGraph, LangSmith, CNN, Transformers,) Tj
-T* (BERT Fine-tuning, RNN, LSTM, GRU, Encoder Decoder, GAN, Pinecone, ChromaDB,) Tj
-T* (MySQL, FastAPI, Docker, Git, Github, Sentence Transformers, Scikit-learn, Llama 4,) Tj
-T* (Gemini API, E5 Multilingual Embeddings, OCR Based Extraction, PyTorch, BERT) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 279.8899999999999864 Td
-(Experience Summary:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 279.8899999999999864 Td
-(AI & Data Science undergraduate with practical experience in architecting and) Tj
-T* (deploying end-to-end AI systems, specializing in Deep Learning, RAG pipelines, and) Tj
-T* (multimodal modeling.) Tj
-ET
-BT
-/F2 16 Tf
-18.3999999999999986 TL
-0 g
-40. 231.8899999999999864 Td
-(Alignment Result) Tj
-ET
-BT
-/F2 11 Tf
-12.6499999999999986 TL
-0 g
-40. 207.8899999999999864 Td
-(Fit Summary:) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-130. 207.8899999999999864 Td
-(The student presents an exceptional fit for the Generative AI Engineer Intern role,) Tj
-T* (showcasing a strong academic foundation in AI/Data Science, practical deployment) Tj
-T* (experience, and highly specialized skills in Generative AI, RAG pipelines, and LLM) Tj
-T* (development. Their demonstrated proficiency in PyTorch and MLOps tools directly) Tj
-T* (aligns with the job's core responsibilities and 'plus' qualifications.) Tj
-ET
-BT
-/F2 12 Tf
-13.7999999999999989 TL
-0 g
-40. 137.8899999999999864 Td
-(Meeting Expectations) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 121.8899999999999864 Td
-(- Strong foundation in Python and deep learning frameworks, specifically PyTorch, aligning with the) Tj
-T* ('Knowledge of deep learning frameworks \(e.g., TensorFlow or PyTorch\) is a plus' requirement.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 93.8899999999999864 Td
-(- Extensive experience with Machine Learning concepts and models, including CNN, Transformers,) Tj
-T* (BERT, RNN, LSTM, GRU, Encoder Decoder, and GANs, demonstrating a robust understanding of AI/ML) Tj
-T* (models and algorithms.) Tj
-ET
-endstream
-endobj
-5 0 obj
-<</Type /Page
-/Parent 1 0 R
-/Resources 2 0 R
-/MediaBox [0 0 595.2799999999999727 841.8899999999999864]
-/Contents 6 0 R
->>
-endobj
-6 0 obj
-<<
-/Length 3330
->>
-stream
-0.200025 w
-0 G
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 795.8899999999999864 Td
-(- Direct and highly relevant skills in Generative AI, RAG Pipelines, Semantic Search, Embedding) Tj
-T* (Models, Vector Similarity Search, and Prompt Engineering, which directly supports 'developing and) Tj
-T* (implementing AI/ML models and algorithms' for a Generative AI role.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 753.8899999999999864 Td
-(- Familiarity with key ML/DL libraries like Scikit-learn, which is explicitly mentioned as a required skill.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 739.8899999999999864 Td
-(- Practical experience with MLOps and deployment tools such as FastAPI, Docker, Git, and Github,) Tj
-T* (indicating the ability to 'integrate AI solutions'.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 711.8899999999999864 Td
-(- Experience with LLMs like Llama 4, Gemini API, and fine-tuning BERT, showing proactive 'research on) Tj
-T* (the latest AI trends and technologies'.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 683.8899999999999864 Td
-(- Skills in managing data for AI, including Pinecone, ChromaDB, MySQL, and OCR Based Extraction,) Tj
-T* (relevant to 'data preprocessing, cleaning, and analysis' and 'model training, evaluation, and optimization'.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 655.8899999999999864 Td
-(- The resume summary highlights 'architecting and deploying end-to-end AI systems', which implies) Tj
-T* (strong analytical and problem-solving skills, as well as the ability to 'collaborate with engineers and) Tj
-T* (product teams'.) Tj
-ET
-BT
-/F2 12 Tf
-13.7999999999999989 TL
-0 g
-40. 607.8899999999999864 Td
-(Missing Expectations) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 591.8899999999999864 Td
-(- While likely used, specific mention of 'NumPy' and 'Pandas' as explicit skills is absent from the resume.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 577.8899999999999864 Td
-(- The resume could more explicitly detail experience in 'data preprocessing, cleaning, and analysis' for) Tj
-T* (diverse datasets, beyond what's implied by 'RAG Pipelines' and 'OCR Based Extraction'.) Tj
-ET
-BT
-/F2 12 Tf
-13.7999999999999989 TL
-0 g
-40. 543.8899999999999864 Td
-(Improvement Suggestions) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 527.8899999999999864 Td
-(- Add 'NumPy' and 'Pandas' to your skills list if you have experience with them, as they are foundational) Tj
-T* (for data manipulation in Python.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 499.8899999999999864 Td
-(- Prepare specific examples from past projects where you handled significant 'data preprocessing,) Tj
-T* (cleaning, and analysis' challenges, detailing the techniques used and the impact.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 471.8899999999999864 Td
-(- When discussing projects, explicitly highlight your contributions to 'documenting experiments,) Tj
-T* (processes, and results' and examples of 'collaborating with engineers and product teams' to showcase) Tj
-T* (teamwork and communication skills.) Tj
-ET
-BT
-/F1 11 Tf
-12.6499999999999986 TL
-0 g
-46. 429.8899999999999864 Td
-(- Quantify your experience where possible \(e.g., 'deployed X RAG pipelines serving Y users', 'improved) Tj
-T* (model performance by Z%'\), to demonstrate impact and scale.) Tj
-ET
-endstream
-endobj
-1 0 obj
-<</Type /Pages
-/Kids [3 0 R 5 0 R ]
-/Count 2
->>
-endobj
-7 0 obj
-<<
-/Type /Font
-/BaseFont /Helvetica
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-8 0 obj
-<<
-/Type /Font
-/BaseFont /Helvetica-Bold
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-9 0 obj
-<<
-/Type /Font
-/BaseFont /Helvetica-Oblique
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-10 0 obj
-<<
-/Type /Font
-/BaseFont /Helvetica-BoldOblique
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-11 0 obj
-<<
-/Type /Font
-/BaseFont /Courier
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-12 0 obj
-<<
-/Type /Font
-/BaseFont /Courier-Bold
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-13 0 obj
-<<
-/Type /Font
-/BaseFont /Courier-Oblique
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-14 0 obj
-<<
-/Type /Font
-/BaseFont /Courier-BoldOblique
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-15 0 obj
-<<
-/Type /Font
-/BaseFont /Times-Roman
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-16 0 obj
-<<
-/Type /Font
-/BaseFont /Times-Bold
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-17 0 obj
-<<
-/Type /Font
-/BaseFont /Times-Italic
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-18 0 obj
-<<
-/Type /Font
-/BaseFont /Times-BoldItalic
-/Subtype /Type1
-/Encoding /WinAnsiEncoding
-/FirstChar 32
-/LastChar 255
->>
-endobj
-19 0 obj
-<<
-/Type /Font
-/BaseFont /ZapfDingbats
-/Subtype /Type1
-/FirstChar 32
-/LastChar 255
->>
-endobj
-20 0 obj
-<<
-/Type /Font
-/BaseFont /Symbol
-/Subtype /Type1
-/FirstChar 32
-/LastChar 255
->>
-endobj
-2 0 obj
-<<
-/ProcSet [/PDF /Text /ImageB /ImageC /ImageI]
-/Font <<
-/F1 7 0 R
-/F2 8 0 R
-/F3 9 0 R
-/F4 10 0 R
-/F5 11 0 R
-/F6 12 0 R
-/F7 13 0 R
-/F8 14 0 R
-/F9 15 0 R
-/F10 16 0 R
-/F11 17 0 R
-/F12 18 0 R
-/F13 19 0 R
-/F14 20 0 R
->>
-/XObject <<
->>
->>
-endobj
-21 0 obj
-<<
-/Producer (jsPDF 4.2.1)
-/CreationDate (D:20260410104544+05'30')
->>
-endobj
-22 0 obj
-<<
-/Type /Catalog
-/Pages 1 0 R
-/OpenAction [3 0 R /FitH null]
-/PageLayout /OneColumn
->>
-endobj
-xref
-0 23
-0000000000 65535 f 
-0000009330 00000 n 
-0000011155 00000 n 
-0000000015 00000 n 
-0000000152 00000 n 
-0000005811 00000 n 
-0000005948 00000 n 
-0000009393 00000 n 
-0000009518 00000 n 
-0000009648 00000 n 
-0000009781 00000 n 
-0000009919 00000 n 
-0000010043 00000 n 
-0000010172 00000 n 
-0000010304 00000 n 
-0000010440 00000 n 
-0000010568 00000 n 
-0000010695 00000 n 
-0000010824 00000 n 
-0000010957 00000 n 
-0000011059 00000 n 
-0000011405 00000 n 
-0000011491 00000 n 
-trailer
-<<
-/Size 23
-/Root 22 0 R
-/Info 21 0 R
-/ID [ <95A654D90B03BE650BD8733007BC1C07> <95A654D90B03BE650BD8733007BC1C07> ]
->>
-startxref
-11595
-%%EOF